hw_score_predict.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 619,
      "metadata": {
        "id": "2PjKIVZZXJK2"
      },
      "outputs": [],
      "source": [
        "import os\n",
        "import re\n",
        "import os\n",
        "import tqdm\n",
        "from glob import glob\n",
        "from pathlib import Path\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "from pprint import pprint\n",
        "import graphviz\n",
        "\n",
        "from collections import defaultdict\n",
        "from bs4 import BeautifulSoup\n",
        "\n",
        "from sklearn.feature_extraction.text import TfidfVectorizer\n",
        "from sklearn.metrics.pairwise import cosine_similarity\n",
        "\n",
        "from sklearn import tree\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.metrics import mean_squared_error, r2_score\n",
        "\n",
        "import seaborn as sns\n",
        "import textstat\n",
        "from textblob import TextBlob\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5JqGHh18tNuS"
      },
      "source": [
        "# 1. Reading the Data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 620,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DmnNAclAXJK3",
        "outputId": "4ca0c49f-eb59-4872-8359-d607a52b8102"
      },
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 127/127 [00:17<00:00,  7.11it/s]\n"
          ]
        }
      ],
      "source": [
        "# Reading the html files\n",
        "THRESHOLD = 20\n",
        "data_path = \"./dataset/*.html\"\n",
        "\n",
        "code2convos = dict()\n",
        "\n",
        "pbar = tqdm.tqdm(sorted(list(glob(data_path))))\n",
        "for path in pbar:\n",
        "    # print(Path.cwd() / path)\n",
        "    file_code = os.path.basename(path).split(\".\")[0]\n",
        "    with open(path, \"r\", encoding=\"latin1\") as fh:\n",
        "\n",
        "        # get the file id to use it as key later on\n",
        "        fid = os.path.basename(path).split(\".\")[0]\n",
        "\n",
        "        # read the html file\n",
        "        html_page = fh.read()\n",
        "\n",
        "        # parse the html file with bs4 so we can extract needed stuff\n",
        "        soup = BeautifulSoup(html_page, \"html.parser\")\n",
        "\n",
        "        # grab the conversations with the data-testid pattern\n",
        "        data_test_id_pattern = re.compile(r\"conversation-turn-[0-9]+\")\n",
        "        conversations = soup.find_all(\"div\", attrs={\"data-testid\": data_test_id_pattern})\n",
        "\n",
        "        convo_texts = []\n",
        "\n",
        "        for i, convo in enumerate(conversations):\n",
        "            convo = convo.find_all(\"div\", attrs={\"data-message-author-role\":re.compile( r\"[user|assistant]\") })\n",
        "            if len(convo) > 0:\n",
        "                role = convo[0].get(\"data-message-author-role\")\n",
        "\n",
        "                # SAVE CODE CELLS\n",
        "                code_cells = convo[0].find_all(\"code\")\n",
        "                all_code_cells = []\n",
        "                for code_cell in code_cells:\n",
        "                    if (len(code_cell.text) > THRESHOLD):\n",
        "                        all_code_cells.append(code_cell.text)\n",
        "\n",
        "                convo_texts.append({\n",
        "                        \"role\" : role,\n",
        "                        \"text\" : convo[0].text,\n",
        "                        \"code\": all_code_cells # save code cells\n",
        "                    }\n",
        "                )\n",
        "\n",
        "        code2convos[file_code] = convo_texts"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 621,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "C2TfEbGWXJK4",
        "outputId": "af42821d-af26-40b7-9f27-4b84aba6d2e4"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "{'code': [],\n",
            " 'role': 'user',\n",
            " 'text': 'Load a CSV file into a Pandas in Python. The file is named '\n",
            "         \"'cs412_hw1_dataset.csv' and contains columns like 'Species', \"\n",
            "         \"'Island', 'Sex', 'Diet', 'Year', 'Life Stage', 'Body Mass (g)', \"\n",
            "         \"'Bill Length (mm)', 'Bill Depth (mm)', 'Flipper Length (mm)', and \"\n",
            "         \"'Health Metrics'. \\n\"}\n"
          ]
        }
      ],
      "source": [
        "#one of the conversations\n",
        "pprint(code2convos[\"0031c86e-81f4-4eef-9e0e-28037abf9883\"][0])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EMGyADoGXJK4"
      },
      "source": [
        "#### Things to do:\n",
        "- Prompt matching with questions\n",
        "- Feature Engineering\n",
        "- Question Grades preparation\n",
        "- Train/Test split\n",
        "- Fitting a model for predicting the scores"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uj20rSIsXJK6"
      },
      "source": [
        "#### Prompt Matching\n",
        "> We want to match the prompts with the questions in the Homework Let's\n",
        "> do it with a simple term frequency vectorizing method. For each prompt,\n",
        "> we will come with a vector that represents it. We will do the same\n",
        "> thing with each of the homework questions. Then, we will calculate the\n",
        "> vectors distanance to do the matching"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 622,
      "metadata": {
        "id": "QMkRhJ38XJK6"
      },
      "outputs": [],
      "source": [
        "prompts = []\n",
        "code2prompts = defaultdict(list)\n",
        "for code , convos in code2convos.items():\n",
        "    user_prompts = []\n",
        "    for conv in convos:\n",
        "        if conv[\"role\"] == \"user\":\n",
        "            prompts.append(conv[\"text\"])\n",
        "            user_prompts.append(conv[\"text\"])\n",
        "    code2prompts[code] = user_prompts"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 623,
      "metadata": {
        "id": "oYAJQrzLXJK6"
      },
      "outputs": [],
      "source": [
        "\n",
        "questions = [\n",
        "    \"\"\"Initialize\n",
        "*   First make a copy of the notebook given to you as a starter.\n",
        "*   Make sure you choose Connect form upper right.\n",
        "*   You may upload the data to the section on your left on Colab, than right click on the .csv file and get the path of the file by clicking on \"Copy Path\". You will be using it when loading the data.\n",
        "\n",
        "\"\"\",\n",
        "\n",
        "#####################\n",
        "    \"\"\"Load training dataset (5 pts)\n",
        "    *  Read the .csv file with the pandas library\n",
        "\"\"\",\n",
        "#####################\n",
        "\"\"\"Understanding the dataset & Preprocessing (15 pts)\n",
        "Understanding the Dataset: (5 pts)\n",
        "> - Find the shape of the dataset (number of samples & number of attributes). (Hint: You can use the **shape** function)\n",
        "> - Display variable names (both dependent and independent).\n",
        "> - Display the summary of the dataset. (Hint: You can use the **info** function)\n",
        "> - Display the first 5 rows from training dataset. (Hint: You can use the **head** function)\n",
        "Preprocessing: (10 pts)\n",
        "\n",
        "> - Check if there are any missing values in the dataset. If there are, you can either drop these values or fill it with most common values in corresponding rows. **Be careful that you have enough data for training the  model.**\n",
        "\n",
        "> - Encode categorical labels with the mappings given in the cell below. (Hint: You can use **map** function)\n",
        "\"\"\",\n",
        "\"\"\"Set X & y, split data (5 pts)\n",
        "\n",
        "*   Shuffle the dataset.\n",
        "*   Seperate your dependent variable X, and your independent variable y. The column health_metrics is y, the rest is X.\n",
        "*   Split training and test sets as 80% and 20%, respectively.\n",
        "\"\"\",\n",
        "#####################\n",
        "\"\"\"Features and Correlations (10 pts)\n",
        "\n",
        "* Correlations of features with health (4 points)\n",
        "Calculate the correlations for all features in dataset. Highlight any strong correlations with the target variable. Plot your results in a heatmap.\n",
        "\n",
        "* Feature Selection (3 points)\n",
        "Select a subset of features that are likely strong predictors, justifying your choices based on the computed correlations.\n",
        "\n",
        "* Hypothetical Driver Features (3 points)\n",
        "Propose two hypothetical features that could enhance the model's predictive accuracy for Y, explaining how they might be derived and their expected impact. Show the resulting correlations with target variable.\n",
        "\n",
        "* __Note:__ You get can get help from GPT.\n",
        "\"\"\",\n",
        "#####################\n",
        "\"\"\"Tune Hyperparameters (20 pts)\n",
        "* Choose 2 hyperparameters to tune. You can use the Scikit learn decision tree documentation for the available hyperparameters *(Hyperparameters are listed under \"Parameters\" in the documentation)*. Use GridSearchCV for hyperparameter tuning, with a cross-validation value of 5. Use validation accuracy to pick the best hyper-parameter values. (15 pts)\n",
        "-Explain the hyperparameters you chose to tune. *(What are the hyperparameters you chose? Why did you choose them?)* (5 pts)\n",
        "\"\"\",\n",
        "#####################\n",
        "\"\"\"Re-train and plot the decision tree with the hyperparameters you have chosen (15 pts)\n",
        "- Re-train model with the hyperparameters you have chosen in part 5). (10 pts)\n",
        "- Plot the tree you have trained. (5 pts)\n",
        "Hint: You can import the **plot_tree** function from the sklearn library.\n",
        "\"\"\",\n",
        "#####################\n",
        "\"\"\"Test your classifier on the test set (20 pts)\n",
        "- Predict the labels of testing data using the tree you have trained in step 6. (10 pts)\n",
        "- Report the classification accuracy. (2 pts)\n",
        "- Plot & investigate the confusion matrix. Fill the following blanks. (8 pts)\n",
        "> The model most frequently mistakes class(es) _________ for class(es) _________.\n",
        "Hint: You can use the confusion_matrix function from sklearn.metrics\n",
        "\"\"\",\n",
        "#####################\n",
        "\"\"\"Find the information gain on the first split (10 pts)\"\"\",\n",
        "#####################\n",
        "]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 624,
      "metadata": {
        "id": "UjXAHygSXJK6"
      },
      "outputs": [],
      "source": [
        "vectorizer = TfidfVectorizer()\n",
        "vectorizer = vectorizer.fit(prompts + questions)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 625,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 256
        },
        "id": "fL36tE0FXJK6",
        "outputId": "b4f75285-0c11-4f7a-c79d-64b386caf942"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>00000</th>\n",
              "      <th>000000</th>\n",
              "      <th>000196</th>\n",
              "      <th>000282</th>\n",
              "      <th>000284</th>\n",
              "      <th>000360</th>\n",
              "      <th>000469</th>\n",
              "      <th>000750</th>\n",
              "      <th>000886</th>\n",
              "      <th>000991</th>\n",
              "      <th>...</th>\n",
              "      <th>yticks</th>\n",
              "      <th>yã</th>\n",
              "      <th>yä</th>\n",
              "      <th>zaman</th>\n",
              "      <th>zdã¼m</th>\n",
              "      <th>zero</th>\n",
              "      <th>zerodivisionerror</th>\n",
              "      <th>zeroth</th>\n",
              "      <th>zip</th>\n",
              "      <th>ã¼nã¼yorum</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>5 rows × 5400 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "   00000  000000  000196  000282  000284  000360  000469  000750  000886  \\\n",
              "0    0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
              "1    0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
              "2    0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
              "3    0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
              "4    0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
              "\n",
              "   000991  ...  yticks   yã   yä  zaman  zdã¼m  zero  zerodivisionerror  \\\n",
              "0     0.0  ...     0.0  0.0  0.0    0.0    0.0   0.0                0.0   \n",
              "1     0.0  ...     0.0  0.0  0.0    0.0    0.0   0.0                0.0   \n",
              "2     0.0  ...     0.0  0.0  0.0    0.0    0.0   0.0                0.0   \n",
              "3     0.0  ...     0.0  0.0  0.0    0.0    0.0   0.0                0.0   \n",
              "4     0.0  ...     0.0  0.0  0.0    0.0    0.0   0.0                0.0   \n",
              "\n",
              "   zeroth  zip  ã¼nã¼yorum  \n",
              "0     0.0  0.0         0.0  \n",
              "1     0.0  0.0         0.0  \n",
              "2     0.0  0.0         0.0  \n",
              "3     0.0  0.0         0.0  \n",
              "4     0.0  0.0         0.0  \n",
              "\n",
              "[5 rows x 5400 columns]"
            ]
          },
          "execution_count": 625,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "questions_TF_IDF = pd.DataFrame(vectorizer.transform(questions).toarray(), columns=vectorizer.get_feature_names_out())\n",
        "questions_TF_IDF.head()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 626,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_XsZZf1wXJK7",
        "outputId": "d70dce4c-6bc2-4c49-c208-eb37b62beb43"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "139235c7-736c-4237-92f0-92e8c116832c.html\n",
            "668ad17e-0240-49f7-b5a7-d22e502554c6.html\n",
            "b0640e51-6879-40cb-a4f5-329f952ef99d.html\n",
            "da6b70d5-29f6-491a-ad46-037c77067128.html\n"
          ]
        }
      ],
      "source": [
        "code2prompts_tf_idf = dict()\n",
        "for code, user_prompts in code2prompts.items():\n",
        "    if len(user_prompts) == 0:\n",
        "        # some files have issues\n",
        "        print(code+\".html\")\n",
        "        continue\n",
        "    prompts_TF_IDF = pd.DataFrame(vectorizer.transform(user_prompts).toarray(), columns=vectorizer.get_feature_names_out())\n",
        "    code2prompts_tf_idf[code] = prompts_TF_IDF"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 627,
      "metadata": {
        "id": "iJnwQnPYXJK7"
      },
      "outputs": [],
      "source": [
        "code2cosine = dict()\n",
        "for code, user_prompts_tf_idf in code2prompts_tf_idf.items():\n",
        "    code2cosine[code] = pd.DataFrame(cosine_similarity(questions_TF_IDF,user_prompts_tf_idf))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 628,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 424
        },
        "id": "_1pisl8QXJK7",
        "outputId": "44d2ff2f-e762-4054-b690-b369cce0bc7f"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>code</th>\n",
              "      <th>Q_0</th>\n",
              "      <th>Q_1</th>\n",
              "      <th>Q_2</th>\n",
              "      <th>Q_3</th>\n",
              "      <th>Q_4</th>\n",
              "      <th>Q_5</th>\n",
              "      <th>Q_6</th>\n",
              "      <th>Q_7</th>\n",
              "      <th>Q_8</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0031c86e-81f4-4eef-9e0e-28037abf9883</td>\n",
              "      <td>0.153541</td>\n",
              "      <td>0.294879</td>\n",
              "      <td>0.522169</td>\n",
              "      <td>0.352243</td>\n",
              "      <td>0.460606</td>\n",
              "      <td>0.553876</td>\n",
              "      <td>0.167250</td>\n",
              "      <td>0.182938</td>\n",
              "      <td>0.195440</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0225686d-b825-4cac-8691-3a3a5343df2b</td>\n",
              "      <td>0.192013</td>\n",
              "      <td>0.795607</td>\n",
              "      <td>0.772184</td>\n",
              "      <td>0.882656</td>\n",
              "      <td>0.607114</td>\n",
              "      <td>0.987511</td>\n",
              "      <td>0.892586</td>\n",
              "      <td>0.570741</td>\n",
              "      <td>0.543866</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>041f950b-c013-409a-a642-cffff60b9d4b</td>\n",
              "      <td>0.258306</td>\n",
              "      <td>0.295923</td>\n",
              "      <td>0.624824</td>\n",
              "      <td>0.351872</td>\n",
              "      <td>0.643038</td>\n",
              "      <td>0.454314</td>\n",
              "      <td>0.540269</td>\n",
              "      <td>0.546506</td>\n",
              "      <td>0.325793</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>04f91058-d0f8-4324-83b2-19c671f433dc</td>\n",
              "      <td>0.145965</td>\n",
              "      <td>0.117841</td>\n",
              "      <td>0.267346</td>\n",
              "      <td>0.316809</td>\n",
              "      <td>0.333889</td>\n",
              "      <td>0.309084</td>\n",
              "      <td>0.192434</td>\n",
              "      <td>0.261892</td>\n",
              "      <td>0.407106</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>089eb66d-4c3a-4f58-b98f-a3774a2efb34</td>\n",
              "      <td>0.344182</td>\n",
              "      <td>0.575528</td>\n",
              "      <td>0.782883</td>\n",
              "      <td>0.624833</td>\n",
              "      <td>0.724872</td>\n",
              "      <td>0.872171</td>\n",
              "      <td>0.684797</td>\n",
              "      <td>0.945305</td>\n",
              "      <td>0.511769</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>118</th>\n",
              "      <td>f24219d6-07f0-4baf-80ac-18475dc5b66f</td>\n",
              "      <td>0.187393</td>\n",
              "      <td>0.209889</td>\n",
              "      <td>0.358875</td>\n",
              "      <td>0.211246</td>\n",
              "      <td>0.149281</td>\n",
              "      <td>0.507128</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>0.862585</td>\n",
              "      <td>0.739634</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>119</th>\n",
              "      <td>f2f18684-4a16-4c05-a2d1-c0f96d1de869</td>\n",
              "      <td>0.144760</td>\n",
              "      <td>0.139495</td>\n",
              "      <td>0.332267</td>\n",
              "      <td>0.950344</td>\n",
              "      <td>0.775698</td>\n",
              "      <td>0.828355</td>\n",
              "      <td>0.374145</td>\n",
              "      <td>0.905688</td>\n",
              "      <td>0.553341</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>120</th>\n",
              "      <td>f852596d-fdca-45aa-9050-d4f76ce6a53c</td>\n",
              "      <td>0.211008</td>\n",
              "      <td>0.908697</td>\n",
              "      <td>0.977725</td>\n",
              "      <td>0.924311</td>\n",
              "      <td>0.888609</td>\n",
              "      <td>0.921306</td>\n",
              "      <td>0.779558</td>\n",
              "      <td>0.767328</td>\n",
              "      <td>0.499508</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>121</th>\n",
              "      <td>f8ec3336-fd48-4654-ad98-62ccfb96d096</td>\n",
              "      <td>0.173215</td>\n",
              "      <td>0.967321</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>0.703987</td>\n",
              "      <td>0.942662</td>\n",
              "      <td>0.907417</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>0.782938</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>122</th>\n",
              "      <td>fb8de815-224c-4d06-9fd4-7156d1a9920d</td>\n",
              "      <td>0.170293</td>\n",
              "      <td>0.712409</td>\n",
              "      <td>0.835236</td>\n",
              "      <td>0.760650</td>\n",
              "      <td>0.895230</td>\n",
              "      <td>0.943543</td>\n",
              "      <td>0.944887</td>\n",
              "      <td>0.844004</td>\n",
              "      <td>0.729632</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>123 rows × 10 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "                                     code       Q_0       Q_1       Q_2  \\\n",
              "0    0031c86e-81f4-4eef-9e0e-28037abf9883  0.153541  0.294879  0.522169   \n",
              "1    0225686d-b825-4cac-8691-3a3a5343df2b  0.192013  0.795607  0.772184   \n",
              "2    041f950b-c013-409a-a642-cffff60b9d4b  0.258306  0.295923  0.624824   \n",
              "3    04f91058-d0f8-4324-83b2-19c671f433dc  0.145965  0.117841  0.267346   \n",
              "4    089eb66d-4c3a-4f58-b98f-a3774a2efb34  0.344182  0.575528  0.782883   \n",
              "..                                    ...       ...       ...       ...   \n",
              "118  f24219d6-07f0-4baf-80ac-18475dc5b66f  0.187393  0.209889  0.358875   \n",
              "119  f2f18684-4a16-4c05-a2d1-c0f96d1de869  0.144760  0.139495  0.332267   \n",
              "120  f852596d-fdca-45aa-9050-d4f76ce6a53c  0.211008  0.908697  0.977725   \n",
              "121  f8ec3336-fd48-4654-ad98-62ccfb96d096  0.173215  0.967321  1.000000   \n",
              "122  fb8de815-224c-4d06-9fd4-7156d1a9920d  0.170293  0.712409  0.835236   \n",
              "\n",
              "          Q_3       Q_4       Q_5       Q_6       Q_7       Q_8  \n",
              "0    0.352243  0.460606  0.553876  0.167250  0.182938  0.195440  \n",
              "1    0.882656  0.607114  0.987511  0.892586  0.570741  0.543866  \n",
              "2    0.351872  0.643038  0.454314  0.540269  0.546506  0.325793  \n",
              "3    0.316809  0.333889  0.309084  0.192434  0.261892  0.407106  \n",
              "4    0.624833  0.724872  0.872171  0.684797  0.945305  0.511769  \n",
              "..        ...       ...       ...       ...       ...       ...  \n",
              "118  0.211246  0.149281  0.507128  1.000000  0.862585  0.739634  \n",
              "119  0.950344  0.775698  0.828355  0.374145  0.905688  0.553341  \n",
              "120  0.924311  0.888609  0.921306  0.779558  0.767328  0.499508  \n",
              "121  0.703987  0.942662  0.907417  1.000000  1.000000  0.782938  \n",
              "122  0.760650  0.895230  0.943543  0.944887  0.844004  0.729632  \n",
              "\n",
              "[123 rows x 10 columns]"
            ]
          },
          "execution_count": 628,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "code2questionmapping = dict()\n",
        "for code, cosine_scores in code2cosine.items():\n",
        "    code2questionmapping[code] = code2cosine[code].max(axis=1).tolist()\n",
        "\n",
        "\n",
        "question_mapping_scores = pd.DataFrame(code2questionmapping).T\n",
        "question_mapping_scores.reset_index(inplace=True)\n",
        "question_mapping_scores.rename(columns={i: f\"Q_{i}\" for i in range(len(questions))}, inplace=True)\n",
        "question_mapping_scores.rename(columns={\"index\" : \"code\"}, inplace=True)\n",
        "\n",
        "question_mapping_scores"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vaY3zdPuXJK7"
      },
      "source": [
        "# 2. Feature Engineering"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uVkLwmFIumDS"
      },
      "source": [
        "## 2.1. Provided Features"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RcrpWfsvuwsC"
      },
      "source": [
        "- Number of prompts that a user asked\n",
        "- Number of complaints that a user makes e.g \"the code gives this error!\"\n",
        "- User prompts average number of characters"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 629,
      "metadata": {
        "id": "F2DaJoe2uzyC"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "139235c7-736c-4237-92f0-92e8c116832c\n",
            "668ad17e-0240-49f7-b5a7-d22e502554c6\n",
            "b0640e51-6879-40cb-a4f5-329f952ef99d\n",
            "da6b70d5-29f6-491a-ad46-037c77067128\n"
          ]
        }
      ],
      "source": [
        "code2features = defaultdict(lambda : defaultdict(int))\n",
        "\n",
        "keywords2search = [\"error\", \"no\", \"thank\", \"next\", \"Entropy\", \"apologize\", \"see\"]\n",
        "keywords2search = [k.lower() for k in keywords2search]\n",
        "\n",
        "for code, convs in code2convos.items():\n",
        "    if len(convs) == 0:\n",
        "        print(code)\n",
        "        continue\n",
        "    for c in convs:\n",
        "        text = c[\"text\"].lower()\n",
        "        if c[\"role\"] == \"user\":\n",
        "            # User Prompts\n",
        "\n",
        "            # count the user prompts\n",
        "            code2features[code][\"#user_prompts\"] += 1\n",
        "\n",
        "            # count the keywords\n",
        "            for kw in keywords2search:\n",
        "                code2features[code][f\"#{kw}\"] +=  len(re.findall(rf\"\\b{kw}\\b\", text))\n",
        "\n",
        "            code2features[code][\"prompt_avg_chars\"] += len(text)\n",
        "        else:\n",
        "            # ChatGPT Responses\n",
        "            code2features[code][\"response_avg_chars\"] += len(text)\n",
        "\n",
        "    # Calculate averages outside the inner loop\n",
        "    if code2features[code][\"#user_prompts\"] > 0:\n",
        "        code2features[code][\"prompt_avg_chars\"] /= code2features[code][\"#user_prompts\"]\n",
        "        code2features[code][\"response_avg_chars\"] /= code2features[code][\"#user_prompts\"]\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 630,
      "metadata": {
        "id": "K_C6yHnxu0F0"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>#user_prompts</th>\n",
              "      <th>#error</th>\n",
              "      <th>#no</th>\n",
              "      <th>#thank</th>\n",
              "      <th>#next</th>\n",
              "      <th>#entropy</th>\n",
              "      <th>#apologize</th>\n",
              "      <th>#see</th>\n",
              "      <th>prompt_avg_chars</th>\n",
              "      <th>response_avg_chars</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0031c86e-81f4-4eef-9e0e-28037abf9883</th>\n",
              "      <td>14.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>407.214286</td>\n",
              "      <td>2218.500000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>0225686d-b825-4cac-8691-3a3a5343df2b</th>\n",
              "      <td>18.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>261.111111</td>\n",
              "      <td>1653.111111</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>041f950b-c013-409a-a642-cffff60b9d4b</th>\n",
              "      <td>9.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>663.000000</td>\n",
              "      <td>1636.222222</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>04f91058-d0f8-4324-83b2-19c671f433dc</th>\n",
              "      <td>20.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>114.100000</td>\n",
              "      <td>1749.900000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>089eb66d-4c3a-4f58-b98f-a3774a2efb34</th>\n",
              "      <td>86.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>26.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>217.081395</td>\n",
              "      <td>1455.441860</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>b61f3f4c-785e-4d7c-b963-4480ece4656f</th>\n",
              "      <td>27.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>8.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>511.148148</td>\n",
              "      <td>1309.185185</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>b73f91f8-732f-4a48-bcbd-eadbbb457a94</th>\n",
              "      <td>17.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>540.529412</td>\n",
              "      <td>1658.588235</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ba18e4e8-2c26-46d4-ba31-cc21947aabd5</th>\n",
              "      <td>27.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>267.740741</td>\n",
              "      <td>1750.518519</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>c0f51763-0c69-45fb-a4d7-6906d53db43d</th>\n",
              "      <td>3.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>69.333333</td>\n",
              "      <td>2732.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>c21be348-17bd-4fdd-88f5-137f6a13cdee</th>\n",
              "      <td>18.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>192.444444</td>\n",
              "      <td>1783.722222</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>100 rows × 10 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "                                      #user_prompts  #error  #no  #thank  \\\n",
              "0031c86e-81f4-4eef-9e0e-28037abf9883           14.0     3.0  1.0     0.0   \n",
              "0225686d-b825-4cac-8691-3a3a5343df2b           18.0     0.0  0.0     0.0   \n",
              "041f950b-c013-409a-a642-cffff60b9d4b            9.0     3.0  0.0     1.0   \n",
              "04f91058-d0f8-4324-83b2-19c671f433dc           20.0     1.0  1.0     0.0   \n",
              "089eb66d-4c3a-4f58-b98f-a3774a2efb34           86.0     1.0  5.0     0.0   \n",
              "...                                             ...     ...  ...     ...   \n",
              "b61f3f4c-785e-4d7c-b963-4480ece4656f           27.0     2.0  0.0     0.0   \n",
              "b73f91f8-732f-4a48-bcbd-eadbbb457a94           17.0     1.0  0.0     0.0   \n",
              "ba18e4e8-2c26-46d4-ba31-cc21947aabd5           27.0     3.0  0.0     0.0   \n",
              "c0f51763-0c69-45fb-a4d7-6906d53db43d            3.0     0.0  0.0     0.0   \n",
              "c21be348-17bd-4fdd-88f5-137f6a13cdee           18.0     3.0  0.0     0.0   \n",
              "\n",
              "                                      #next  #entropy  #apologize  #see  \\\n",
              "0031c86e-81f4-4eef-9e0e-28037abf9883    0.0       0.0         0.0   0.0   \n",
              "0225686d-b825-4cac-8691-3a3a5343df2b    0.0       3.0         0.0   0.0   \n",
              "041f950b-c013-409a-a642-cffff60b9d4b    0.0       3.0         0.0   0.0   \n",
              "04f91058-d0f8-4324-83b2-19c671f433dc    0.0       3.0         0.0   1.0   \n",
              "089eb66d-4c3a-4f58-b98f-a3774a2efb34    0.0      26.0         0.0   3.0   \n",
              "...                                     ...       ...         ...   ...   \n",
              "b61f3f4c-785e-4d7c-b963-4480ece4656f    0.0       8.0         0.0   0.0   \n",
              "b73f91f8-732f-4a48-bcbd-eadbbb457a94    0.0       0.0         0.0   1.0   \n",
              "ba18e4e8-2c26-46d4-ba31-cc21947aabd5    4.0       0.0         0.0   0.0   \n",
              "c0f51763-0c69-45fb-a4d7-6906d53db43d    0.0       0.0         0.0   0.0   \n",
              "c21be348-17bd-4fdd-88f5-137f6a13cdee    0.0       4.0         0.0   0.0   \n",
              "\n",
              "                                      prompt_avg_chars  response_avg_chars  \n",
              "0031c86e-81f4-4eef-9e0e-28037abf9883        407.214286         2218.500000  \n",
              "0225686d-b825-4cac-8691-3a3a5343df2b        261.111111         1653.111111  \n",
              "041f950b-c013-409a-a642-cffff60b9d4b        663.000000         1636.222222  \n",
              "04f91058-d0f8-4324-83b2-19c671f433dc        114.100000         1749.900000  \n",
              "089eb66d-4c3a-4f58-b98f-a3774a2efb34        217.081395         1455.441860  \n",
              "...                                                ...                 ...  \n",
              "b61f3f4c-785e-4d7c-b963-4480ece4656f        511.148148         1309.185185  \n",
              "b73f91f8-732f-4a48-bcbd-eadbbb457a94        540.529412         1658.588235  \n",
              "ba18e4e8-2c26-46d4-ba31-cc21947aabd5        267.740741         1750.518519  \n",
              "c0f51763-0c69-45fb-a4d7-6906d53db43d         69.333333         2732.000000  \n",
              "c21be348-17bd-4fdd-88f5-137f6a13cdee        192.444444         1783.722222  \n",
              "\n",
              "[100 rows x 10 columns]"
            ]
          },
          "execution_count": 630,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "df = pd.DataFrame(code2features).T\n",
        "df.head(100)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ryz_bnmtp2H1"
      },
      "source": [
        "## 2.2. Custom Features"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vEUofgQXqSeh"
      },
      "source": [
        "### a. Number of Apologize Words"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 631,
      "metadata": {
        "id": "SRgcYElw-oxQ"
      },
      "outputs": [],
      "source": [
        "from collections import defaultdict\n",
        "import re\n",
        "\n",
        "def count_confusion_and_apologize_in_sentence(text):\n",
        "    # Splitting the text into sentences\n",
        "    sentences = re.split(r'[.!?]', text)\n",
        "\n",
        "    # Counting sentences that contain both \"confusion\" and \"apologize\"\n",
        "    count = 0\n",
        "    for sentence in sentences:\n",
        "        # Check both words in the lowercased sentence\n",
        "        if 'confusion' in sentence.lower() and 'apologize' in sentence.lower():\n",
        "            count += 1\n",
        "\n",
        "    return count"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Bu-Nz6tBqZ2u"
      },
      "source": [
        "### b. Number of Code Lines & Cells"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 632,
      "metadata": {
        "id": "r5FGqz77pcs1"
      },
      "outputs": [],
      "source": [
        "# CUSTOM FEATURES 2: NUMBER OF CODE LINES AND NUMBER OF CODE CELLS PROVIDED BY CHATGPT FOR EACH QUESTION\n",
        "code2num_code_lines = {}\n",
        "code2num_code_cells = {}\n",
        "\n",
        "# Format\n",
        "# {\n",
        "#   code_1: {\n",
        "#       Q1: 12,\n",
        "#       Q2: 20,\n",
        "#   }\n",
        "# }\n",
        "def DICT_INIT(feature_name):\n",
        "    _dict = {\n",
        "        f\"Q_0_{feature_name}\": 0,\n",
        "        f\"Q_1_{feature_name}\": 0,\n",
        "        f\"Q_2_{feature_name}\": 0,\n",
        "        f\"Q_3_{feature_name}\": 0,\n",
        "        f\"Q_4_{feature_name}\": 0,\n",
        "        f\"Q_5_{feature_name}\": 0,\n",
        "        f\"Q_6_{feature_name}\": 0,\n",
        "        f\"Q_7_{feature_name}\": 0,\n",
        "        f\"Q_8_{feature_name}\": 0,\n",
        "    }\n",
        "    return _dict\n",
        "\n",
        "FEATURE_NUM_CODE_LINES = \"#code_lines\"\n",
        "FEATURE_NUM_CODE_CELLS = \"#code_cells\"\n",
        "\n",
        "# Iterate over all file to cosine similarities\n",
        "for file_code in code2cosine:\n",
        "    # Initialize file_code entries\n",
        "    code2num_code_lines[file_code] = DICT_INIT(FEATURE_NUM_CODE_LINES)\n",
        "    code2num_code_cells[file_code] = DICT_INIT(FEATURE_NUM_CODE_CELLS)\n",
        "    for prompt_id in code2cosine[file_code]:\n",
        "        similarities = code2cosine[file_code][prompt_id]\n",
        "        # Find the max similartity to relate the prompt with a question index\n",
        "        related_question_idx = similarities.idxmax()\n",
        "\n",
        "        # Get the GPT answer to count the number of code cells and lines\n",
        "        user_prompt_idx = -1\n",
        "        found = False\n",
        "        # Find the user prompt index in all conversations\n",
        "        for conv in code2convos[file_code]:\n",
        "            if conv[\"role\"] == \"user\":\n",
        "                user_prompt_idx += 1\n",
        "                if (user_prompt_idx == prompt_id):\n",
        "                    found = True\n",
        "            # If found and role is not user, set this prompt as the GPT answer index\n",
        "            elif found:\n",
        "                break\n",
        "\n",
        "\n",
        "        code_cells = conv[\"code\"]\n",
        "        for code_cell in code_cells:\n",
        "            # Increment num code cells of the related question by one\n",
        "            code2num_code_cells[file_code][f\"Q_{related_question_idx}_{FEATURE_NUM_CODE_CELLS}\"] += 1\n",
        "            # Increment num code lines of the related question by the '\\n' count\n",
        "            code2num_code_lines[file_code][f\"Q_{related_question_idx}_{FEATURE_NUM_CODE_LINES}\"] += len(code_cell.split(\"\\n\"))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 633,
      "metadata": {
        "id": "koZKfEWFpwQq"
      },
      "outputs": [],
      "source": [
        "# Create dataframes out of generated dictionaries and merge them\n",
        "code_lines_df = (pd.DataFrame(code2num_code_lines)).T\n",
        "code_lines_df = code_lines_df.reset_index().rename(columns={\"index\": \"code\"})\n",
        "\n",
        "code_cells_df = (pd.DataFrame(code2num_code_cells)).T\n",
        "code_cells_df = code_cells_df.reset_index().rename(columns={\"index\": \"code\"})\n",
        "\n",
        "code_df = pd.merge(code_lines_df, code_cells_df, on='code', how='inner')\n",
        "code_df.to_csv(\"./generated_data/code_features.csv\", index=False)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "### c. Prompt Complexity Analysis"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 634,
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "                                   code  average_fk_grade\n",
            "0  0031c86e-81f4-4eef-9e0e-28037abf9883         12.078571\n",
            "1  0225686d-b825-4cac-8691-3a3a5343df2b         12.872222\n",
            "2  041f950b-c013-409a-a642-cffff60b9d4b         14.866667\n",
            "3  04f91058-d0f8-4324-83b2-19c671f433dc         11.030000\n",
            "4  089eb66d-4c3a-4f58-b98f-a3774a2efb34         13.123256\n"
          ]
        }
      ],
      "source": [
        "# Function to calculate the average Flesch-Kincaid Grade Level for ChatGPT responses in a conversation\n",
        "def calculate_average_fk_grade(responses):\n",
        "    total_fk_grade = 0\n",
        "    count = 0\n",
        "    for response in responses:\n",
        "        fk_grade = textstat.flesch_kincaid_grade(response)\n",
        "        total_fk_grade += fk_grade\n",
        "        count += 1\n",
        "    return total_fk_grade / count if count > 0 else 0\n",
        "\n",
        "for code, convos in code2convos.items():\n",
        "    assistant_responses = [convo['text'] for convo in convos if convo['role'] == 'assistant']\n",
        "    avg_fk_grade = calculate_average_fk_grade(assistant_responses)\n",
        "    code2features[code]['avg_fk_grade'] = avg_fk_grade\n",
        "\n",
        "# Assuming code2features is your existing dictionary with the calculated avg_fk_grade\n",
        "# Convert code2features to a DataFrame\n",
        "features_df = pd.DataFrame.from_dict(code2features, orient='index').reset_index()\n",
        "features_df.rename(columns={'index': 'code', 'avg_fk_grade': 'average_fk_grade'}, inplace=True)\n",
        "\n",
        "# Extract only the 'code' and 'average_fk_grade' columns\n",
        "fk_grade_df = features_df[['code', 'average_fk_grade']]\n",
        "\n",
        "# Display the DataFrame\n",
        "print(fk_grade_df.head())"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HklW_rk1rh9M"
      },
      "source": [
        "### d. Average Prompt per Question"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 635,
      "metadata": {
        "id": "ml2E0sg7rljg"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>code</th>\n",
              "      <th>Q_0_#prompts</th>\n",
              "      <th>Q_1_#prompts</th>\n",
              "      <th>Q_2_#prompts</th>\n",
              "      <th>Q_3_#prompts</th>\n",
              "      <th>Q_4_#prompts</th>\n",
              "      <th>Q_5_#prompts</th>\n",
              "      <th>Q_6_#prompts</th>\n",
              "      <th>Q_7_#prompts</th>\n",
              "      <th>Q_8_#prompts</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0031c86e-81f4-4eef-9e0e-28037abf9883</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>6.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0225686d-b825-4cac-8691-3a3a5343df2b</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>041f950b-c013-409a-a642-cffff60b9d4b</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>04f91058-d0f8-4324-83b2-19c671f433dc</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>6.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>6.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>089eb66d-4c3a-4f58-b98f-a3774a2efb34</td>\n",
              "      <td>5.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>19.0</td>\n",
              "      <td>15.0</td>\n",
              "      <td>17.0</td>\n",
              "      <td>9.0</td>\n",
              "      <td>7.0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>9.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>118</th>\n",
              "      <td>f24219d6-07f0-4baf-80ac-18475dc5b66f</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>3.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>119</th>\n",
              "      <td>f2f18684-4a16-4c05-a2d1-c0f96d1de869</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>35.0</td>\n",
              "      <td>7.0</td>\n",
              "      <td>7.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>6.0</td>\n",
              "      <td>6.0</td>\n",
              "      <td>5.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>120</th>\n",
              "      <td>f852596d-fdca-45aa-9050-d4f76ce6a53c</td>\n",
              "      <td>1.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>7.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>7.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>2.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>121</th>\n",
              "      <td>f8ec3336-fd48-4654-ad98-62ccfb96d096</td>\n",
              "      <td>2.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>122</th>\n",
              "      <td>fb8de815-224c-4d06-9fd4-7156d1a9920d</td>\n",
              "      <td>3.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>7.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>6.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>3.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>123 rows × 10 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "                                     code  Q_0_#prompts  Q_1_#prompts  \\\n",
              "0    0031c86e-81f4-4eef-9e0e-28037abf9883           1.0           1.0   \n",
              "1    0225686d-b825-4cac-8691-3a3a5343df2b           0.0           1.0   \n",
              "2    041f950b-c013-409a-a642-cffff60b9d4b           0.0           0.0   \n",
              "3    04f91058-d0f8-4324-83b2-19c671f433dc           0.0           0.0   \n",
              "4    089eb66d-4c3a-4f58-b98f-a3774a2efb34           5.0           1.0   \n",
              "..                                    ...           ...           ...   \n",
              "118  f24219d6-07f0-4baf-80ac-18475dc5b66f           0.0           1.0   \n",
              "119  f2f18684-4a16-4c05-a2d1-c0f96d1de869           0.0           0.0   \n",
              "120  f852596d-fdca-45aa-9050-d4f76ce6a53c           1.0           2.0   \n",
              "121  f8ec3336-fd48-4654-ad98-62ccfb96d096           2.0           1.0   \n",
              "122  fb8de815-224c-4d06-9fd4-7156d1a9920d           3.0           1.0   \n",
              "\n",
              "     Q_2_#prompts  Q_3_#prompts  Q_4_#prompts  Q_5_#prompts  Q_6_#prompts  \\\n",
              "0             2.0           3.0           6.0           1.0           0.0   \n",
              "1             5.0           1.0           4.0           2.0           1.0   \n",
              "2             2.0           3.0           1.0           0.0           2.0   \n",
              "3             6.0           3.0           6.0           1.0           1.0   \n",
              "4            19.0          15.0          17.0           9.0           7.0   \n",
              "..            ...           ...           ...           ...           ...   \n",
              "118           5.0           3.0           2.0           2.0           2.0   \n",
              "119          35.0           7.0           7.0           5.0           6.0   \n",
              "120           7.0           3.0           7.0           2.0           2.0   \n",
              "121           2.0           2.0           2.0           2.0           1.0   \n",
              "122           7.0           2.0           5.0           1.0           6.0   \n",
              "\n",
              "     Q_7_#prompts  Q_8_#prompts  \n",
              "0             0.0           0.0  \n",
              "1             3.0           1.0  \n",
              "2             1.0           0.0  \n",
              "3             2.0           1.0  \n",
              "4             4.0           9.0  \n",
              "..            ...           ...  \n",
              "118           3.0           3.0  \n",
              "119           6.0           5.0  \n",
              "120           4.0           2.0  \n",
              "121           1.0           1.0  \n",
              "122           3.0           3.0  \n",
              "\n",
              "[123 rows x 10 columns]"
            ]
          },
          "execution_count": 635,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "rename_map = {}\n",
        "for i in range(9):\n",
        "    rename_map[i] = f\"Q_{i}_#prompts\"\n",
        "\n",
        "# Average Prompt per Question\n",
        "def promptCount(code2cosine):\n",
        "    # most similar question for each prompt\n",
        "    code2mostSimilarQuestion = dict()\n",
        "    for code, cosine_scores in code2cosine.items():\n",
        "        # cosine similarity\n",
        "        mostSimilarQuestions = cosine_scores.idxmax(axis=0)\n",
        "        code2mostSimilarQuestion[code] = mostSimilarQuestions\n",
        "\n",
        "    # number of prompts for each question\n",
        "    promptCountsPerQuestion = dict()\n",
        "    for code, questionIndices in code2mostSimilarQuestion.items():\n",
        "        promptCounts = questionIndices.value_counts().sort_index()\n",
        "        promptCountsPerQuestion[code] = promptCounts\n",
        "\n",
        "    promptCountsDF = pd.DataFrame(promptCountsPerQuestion).T\n",
        "    promptCountsDF.reset_index(inplace=True)\n",
        "    promptCountsDF.rename(columns={\"index\": \"code\"}, inplace=True)\n",
        "    promptCountsDF.rename(columns=rename_map, inplace=True)\n",
        "    promptCountsDF.fillna(0, inplace=True)\n",
        "\n",
        "    # average prompts per question\n",
        "    averagePromptCounts = promptCountsDF.drop('code', axis=1).mean()\n",
        "    averagePromptCounts.name = 'average_prompt_count_per_question'\n",
        "\n",
        "    return promptCountsDF, averagePromptCounts\n",
        "\n",
        "\n",
        "promptCountsDF, averagePromptCounts = promptCount(code2cosine)\n",
        "sum_last_8 = promptCountsDF.iloc[:, -8:].sum(axis=1).tolist()\n",
        "prompt_ratio = [value / sum(averagePromptCounts) / len(averagePromptCounts) for value in sum_last_8]\n",
        "\n",
        "zeros_count_per_row = (promptCountsDF == 0).sum(axis=1).tolist()\n",
        "zeros_count_per_row = [count * -1 for count in zeros_count_per_row]\n",
        "\n",
        "promptCountsDF"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "### e. Sentiment Analysis"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 636,
      "metadata": {},
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "/Library/Frameworks/Python.framework/Versions/3.10/lib/python3.10/site-packages/numpy/lib/function_base.py:518: RuntimeWarning: Mean of empty slice.\n",
            "  avg = a.mean(axis, **keepdims_kw)\n",
            "/Library/Frameworks/Python.framework/Versions/3.10/lib/python3.10/site-packages/numpy/core/_methods.py:190: RuntimeWarning: invalid value encountered in double_scalars\n",
            "  ret = ret.dtype.type(ret / rcount)\n"
          ]
        },
        {
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>code</th>\n",
              "      <th>average_sentiment</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0031c86e-81f4-4eef-9e0e-28037abf9883</td>\n",
              "      <td>0.139457</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0225686d-b825-4cac-8691-3a3a5343df2b</td>\n",
              "      <td>0.130718</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>041f950b-c013-409a-a642-cffff60b9d4b</td>\n",
              "      <td>-0.021337</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>04f91058-d0f8-4324-83b2-19c671f433dc</td>\n",
              "      <td>0.004042</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>089eb66d-4c3a-4f58-b98f-a3774a2efb34</td>\n",
              "      <td>0.089507</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                   code  average_sentiment\n",
              "0  0031c86e-81f4-4eef-9e0e-28037abf9883           0.139457\n",
              "1  0225686d-b825-4cac-8691-3a3a5343df2b           0.130718\n",
              "2  041f950b-c013-409a-a642-cffff60b9d4b          -0.021337\n",
              "3  04f91058-d0f8-4324-83b2-19c671f433dc           0.004042\n",
              "4  089eb66d-4c3a-4f58-b98f-a3774a2efb34           0.089507"
            ]
          },
          "execution_count": 636,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# Analyze sentiment of each part\n",
        "codes = []\n",
        "average_sentiments = []\n",
        "sentiment_df = pd.DataFrame()\n",
        "\n",
        "for code in code2prompts:\n",
        "    prompts = code2prompts[code]\n",
        "    sentiments = [TextBlob(text).sentiment.polarity for text in prompts]\n",
        "    average_sentiment = np.average(sentiments)\n",
        "    if (np.isnan(average_sentiment)):\n",
        "        average_sentiment = 0\n",
        "    codes.append(code)\n",
        "    average_sentiments.append(average_sentiment)\n",
        "\n",
        "sentiment_df[\"code\"] = codes\n",
        "sentiment_df[\"average_sentiment\"] = average_sentiments\n",
        "sentiment_df.fillna(0)\n",
        "\n",
        "sentiment_df.head()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "### f. Message Length Variability"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 637,
      "metadata": {},
      "outputs": [
        {
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>code</th>\n",
              "      <th>message_length_variability</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0031c86e-81f4-4eef-9e0e-28037abf9883</td>\n",
              "      <td>22.901854</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0225686d-b825-4cac-8691-3a3a5343df2b</td>\n",
              "      <td>33.739424</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>041f950b-c013-409a-a642-cffff60b9d4b</td>\n",
              "      <td>229.512097</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>04f91058-d0f8-4324-83b2-19c671f433dc</td>\n",
              "      <td>11.188722</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>089eb66d-4c3a-4f58-b98f-a3774a2efb34</td>\n",
              "      <td>21.204801</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                   code  message_length_variability\n",
              "0  0031c86e-81f4-4eef-9e0e-28037abf9883                   22.901854\n",
              "1  0225686d-b825-4cac-8691-3a3a5343df2b                   33.739424\n",
              "2  041f950b-c013-409a-a642-cffff60b9d4b                  229.512097\n",
              "3  04f91058-d0f8-4324-83b2-19c671f433dc                   11.188722\n",
              "4  089eb66d-4c3a-4f58-b98f-a3774a2efb34                   21.204801"
            ]
          },
          "execution_count": 637,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "def calculate_length_variability(conversation):\n",
        "    user_message_lengths = [len(message['text'].split()) for message in conversation if message['role'] == 'user']\n",
        "    if user_message_lengths:\n",
        "        return np.std(user_message_lengths)\n",
        "    else:\n",
        "        return 0\n",
        "\n",
        "# Calculating message length variability for each user\n",
        "user_message_variability = {code: calculate_length_variability(convos) for code, convos in code2convos.items()}\n",
        "\n",
        "# Convert to DataFrame\n",
        "message_variability_df = pd.DataFrame.from_dict(user_message_variability, orient='index', columns=['message_length_variability'])\n",
        "\n",
        "# Reset index to obtain 'code' as a column\n",
        "message_variability_df.reset_index(inplace=True)\n",
        "message_variability_df.rename(columns={'index': 'code'}, inplace=True)\n",
        "\n",
        "# Display the DataFrame\n",
        "message_variability_df.head()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "### g. Engagement Score "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 638,
      "metadata": {},
      "outputs": [
        {
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>code</th>\n",
              "      <th>engagement_score</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0031c86e-81f4-4eef-9e0e-28037abf9883</td>\n",
              "      <td>97.517857</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0225686d-b825-4cac-8691-3a3a5343df2b</td>\n",
              "      <td>83.347222</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>041f950b-c013-409a-a642-cffff60b9d4b</td>\n",
              "      <td>109.138889</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>04f91058-d0f8-4324-83b2-19c671f433dc</td>\n",
              "      <td>55.437500</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>089eb66d-4c3a-4f58-b98f-a3774a2efb34</td>\n",
              "      <td>209.691860</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>122</th>\n",
              "      <td>f24219d6-07f0-4baf-80ac-18475dc5b66f</td>\n",
              "      <td>79.214286</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>123</th>\n",
              "      <td>f2f18684-4a16-4c05-a2d1-c0f96d1de869</td>\n",
              "      <td>246.940141</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>124</th>\n",
              "      <td>f852596d-fdca-45aa-9050-d4f76ce6a53c</td>\n",
              "      <td>178.858333</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>125</th>\n",
              "      <td>f8ec3336-fd48-4654-ad98-62ccfb96d096</td>\n",
              "      <td>115.214286</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>126</th>\n",
              "      <td>fb8de815-224c-4d06-9fd4-7156d1a9920d</td>\n",
              "      <td>153.653226</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>127 rows × 2 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "                                     code  engagement_score\n",
              "0    0031c86e-81f4-4eef-9e0e-28037abf9883         97.517857\n",
              "1    0225686d-b825-4cac-8691-3a3a5343df2b         83.347222\n",
              "2    041f950b-c013-409a-a642-cffff60b9d4b        109.138889\n",
              "3    04f91058-d0f8-4324-83b2-19c671f433dc         55.437500\n",
              "4    089eb66d-4c3a-4f58-b98f-a3774a2efb34        209.691860\n",
              "..                                    ...               ...\n",
              "122  f24219d6-07f0-4baf-80ac-18475dc5b66f         79.214286\n",
              "123  f2f18684-4a16-4c05-a2d1-c0f96d1de869        246.940141\n",
              "124  f852596d-fdca-45aa-9050-d4f76ce6a53c        178.858333\n",
              "125  f8ec3336-fd48-4654-ad98-62ccfb96d096        115.214286\n",
              "126  fb8de815-224c-4d06-9fd4-7156d1a9920d        153.653226\n",
              "\n",
              "[127 rows x 2 columns]"
            ]
          },
          "execution_count": 638,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "from collections import Counter\n",
        "import numpy as np\n",
        "import re\n",
        "\n",
        "# Function to calculate topic diversity\n",
        "def calculate_topic_diversity(messages):\n",
        "    words = [word.lower() for message in messages for word in re.findall(r'\\b\\w+\\b', message['text'])]\n",
        "    unique_words = Counter(words)\n",
        "    topic_diversity = len(unique_words)\n",
        "    return topic_diversity\n",
        "\n",
        "# Function to calculate user engagement score\n",
        "def calculate_user_engagement_score(conversation):\n",
        "    user_messages = [message for message in conversation if message['role'] == 'user']\n",
        "    chatgpt_responses = [message for message in conversation if message['role'] == 'assistant']\n",
        "\n",
        "    message_frequency = len(user_messages)\n",
        "    topic_diversity = calculate_topic_diversity(user_messages)\n",
        "    avg_message_length = np.mean([len(message['text'].split()) for message in user_messages]) if user_messages else 0\n",
        "    responsiveness_score = len(user_messages) / len(chatgpt_responses) if chatgpt_responses else 0\n",
        "\n",
        "    engagement_score = (message_frequency + topic_diversity + avg_message_length + responsiveness_score) / 4\n",
        "    return engagement_score\n",
        "\n",
        "# Assuming 'code2convos' contains the conversation data for each user\n",
        "# Example: code2convos = {'user_code1': conversation_data1, 'user_code2': conversation_data2, ...}\n",
        "\n",
        "# Calculate user engagement score for each user\n",
        "user_engagement_scores = {code: calculate_user_engagement_score(convos) for code, convos in code2convos.items()}\n",
        "\n",
        "# Convert to DataFrame\n",
        "engagement_scores_df = pd.DataFrame.from_dict(user_engagement_scores, orient='index', columns=['engagement_score'])\n",
        "\n",
        "# Reset index to obtain 'code' as a column\n",
        "engagement_scores_df.reset_index(inplace=True)\n",
        "engagement_scores_df.rename(columns={'index': 'code'}, inplace=True)\n",
        "\n",
        "# Display the DataFrame\n",
        "engagement_scores_df"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5U--jc1JrtIO"
      },
      "source": [
        "## 2.3. Merging Features and Target Variable"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 639,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "8Vy4RxroXJK8",
        "outputId": "6742d956-e85b-4915-acf6-8dc3ea47a8ed"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>code</th>\n",
              "      <th>grade</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>6a2003ad-a05a-41c9-9d48-e98491a90499</td>\n",
              "      <td>90.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>04f91058-d0f8-4324-83b2-19c671f433dc</td>\n",
              "      <td>97.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>81fdeb2a-e7e5-4a05-8058-d31ea579b0d9</td>\n",
              "      <td>94.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>6a903495-c5be-4263-b4dd-75e2bbc30434</td>\n",
              "      <td>97.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>6d5742c1-77c4-429c-8f6e-ef1262ca5557</td>\n",
              "      <td>93.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                   code  grade\n",
              "0  6a2003ad-a05a-41c9-9d48-e98491a90499   90.0\n",
              "1  04f91058-d0f8-4324-83b2-19c671f433dc   97.0\n",
              "2  81fdeb2a-e7e5-4a05-8058-d31ea579b0d9   94.0\n",
              "3  6a903495-c5be-4263-b4dd-75e2bbc30434   97.0\n",
              "4  6d5742c1-77c4-429c-8f6e-ef1262ca5557   93.0"
            ]
          },
          "execution_count": 639,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# reading the scores\n",
        "scores = pd.read_csv(\"./project-material/scores.csv\", sep=\",\")\n",
        "scores[\"code\"] = scores[\"code\"].apply(lambda x: x.strip())\n",
        "\n",
        "# selecting the columns we need and we care\n",
        "scores = scores[[\"code\", \"grade\"]]\n",
        "\n",
        "# show some examples\n",
        "scores.head()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 640,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 487
        },
        "id": "aqg-th0NXJK8",
        "outputId": "f4ee056a-17cc-4e70-c048-b92914b29149"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>code</th>\n",
              "      <th>#user_prompts</th>\n",
              "      <th>#error</th>\n",
              "      <th>#no</th>\n",
              "      <th>#thank</th>\n",
              "      <th>#next</th>\n",
              "      <th>#entropy</th>\n",
              "      <th>#apologize</th>\n",
              "      <th>#see</th>\n",
              "      <th>prompt_avg_chars</th>\n",
              "      <th>response_avg_chars</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0031c86e-81f4-4eef-9e0e-28037abf9883</td>\n",
              "      <td>14.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>407.214286</td>\n",
              "      <td>2218.500000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0225686d-b825-4cac-8691-3a3a5343df2b</td>\n",
              "      <td>18.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>261.111111</td>\n",
              "      <td>1653.111111</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>041f950b-c013-409a-a642-cffff60b9d4b</td>\n",
              "      <td>9.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>663.000000</td>\n",
              "      <td>1636.222222</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>04f91058-d0f8-4324-83b2-19c671f433dc</td>\n",
              "      <td>20.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>114.100000</td>\n",
              "      <td>1749.900000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>089eb66d-4c3a-4f58-b98f-a3774a2efb34</td>\n",
              "      <td>86.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>26.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>217.081395</td>\n",
              "      <td>1455.441860</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                   code  #user_prompts  #error  #no  #thank  \\\n",
              "0  0031c86e-81f4-4eef-9e0e-28037abf9883           14.0     3.0  1.0     0.0   \n",
              "1  0225686d-b825-4cac-8691-3a3a5343df2b           18.0     0.0  0.0     0.0   \n",
              "2  041f950b-c013-409a-a642-cffff60b9d4b            9.0     3.0  0.0     1.0   \n",
              "3  04f91058-d0f8-4324-83b2-19c671f433dc           20.0     1.0  1.0     0.0   \n",
              "4  089eb66d-4c3a-4f58-b98f-a3774a2efb34           86.0     1.0  5.0     0.0   \n",
              "\n",
              "   #next  #entropy  #apologize  #see  prompt_avg_chars  response_avg_chars  \n",
              "0    0.0       0.0         0.0   0.0        407.214286         2218.500000  \n",
              "1    0.0       3.0         0.0   0.0        261.111111         1653.111111  \n",
              "2    0.0       3.0         0.0   0.0        663.000000         1636.222222  \n",
              "3    0.0       3.0         0.0   1.0        114.100000         1749.900000  \n",
              "4    0.0      26.0         0.0   3.0        217.081395         1455.441860  "
            ]
          },
          "execution_count": 640,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "df.reset_index(inplace=True, drop=False)\n",
        "df.rename(columns={\"index\": \"code\"}, inplace=True)\n",
        "df.head()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 641,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 487
        },
        "id": "kh28QER-XJK8",
        "outputId": "272ef4f6-1a30-4bfd-ac29-5ae91bccb61e"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>code</th>\n",
              "      <th>#user_prompts</th>\n",
              "      <th>prompt_avg_chars</th>\n",
              "      <th>response_avg_chars</th>\n",
              "      <th>Q_0</th>\n",
              "      <th>Q_1</th>\n",
              "      <th>Q_2</th>\n",
              "      <th>Q_3</th>\n",
              "      <th>Q_4</th>\n",
              "      <th>Q_5</th>\n",
              "      <th>...</th>\n",
              "      <th>Q_1_#prompts</th>\n",
              "      <th>Q_2_#prompts</th>\n",
              "      <th>Q_3_#prompts</th>\n",
              "      <th>Q_4_#prompts</th>\n",
              "      <th>Q_5_#prompts</th>\n",
              "      <th>Q_6_#prompts</th>\n",
              "      <th>Q_7_#prompts</th>\n",
              "      <th>Q_8_#prompts</th>\n",
              "      <th>average_fk_grade</th>\n",
              "      <th>grade</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0031c86e-81f4-4eef-9e0e-28037abf9883</td>\n",
              "      <td>14.0</td>\n",
              "      <td>407.214286</td>\n",
              "      <td>2218.500000</td>\n",
              "      <td>0.153541</td>\n",
              "      <td>0.294879</td>\n",
              "      <td>0.522169</td>\n",
              "      <td>0.352243</td>\n",
              "      <td>0.460606</td>\n",
              "      <td>0.553876</td>\n",
              "      <td>...</td>\n",
              "      <td>1.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>6.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>12.078571</td>\n",
              "      <td>48.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0225686d-b825-4cac-8691-3a3a5343df2b</td>\n",
              "      <td>18.0</td>\n",
              "      <td>261.111111</td>\n",
              "      <td>1653.111111</td>\n",
              "      <td>0.192013</td>\n",
              "      <td>0.795607</td>\n",
              "      <td>0.772184</td>\n",
              "      <td>0.882656</td>\n",
              "      <td>0.607114</td>\n",
              "      <td>0.987511</td>\n",
              "      <td>...</td>\n",
              "      <td>1.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>12.872222</td>\n",
              "      <td>99.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>041f950b-c013-409a-a642-cffff60b9d4b</td>\n",
              "      <td>9.0</td>\n",
              "      <td>663.000000</td>\n",
              "      <td>1636.222222</td>\n",
              "      <td>0.258306</td>\n",
              "      <td>0.295923</td>\n",
              "      <td>0.624824</td>\n",
              "      <td>0.351872</td>\n",
              "      <td>0.643038</td>\n",
              "      <td>0.454314</td>\n",
              "      <td>...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>14.866667</td>\n",
              "      <td>90.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>04f91058-d0f8-4324-83b2-19c671f433dc</td>\n",
              "      <td>20.0</td>\n",
              "      <td>114.100000</td>\n",
              "      <td>1749.900000</td>\n",
              "      <td>0.145965</td>\n",
              "      <td>0.117841</td>\n",
              "      <td>0.267346</td>\n",
              "      <td>0.316809</td>\n",
              "      <td>0.333889</td>\n",
              "      <td>0.309084</td>\n",
              "      <td>...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>6.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>6.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>11.030000</td>\n",
              "      <td>97.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>089eb66d-4c3a-4f58-b98f-a3774a2efb34</td>\n",
              "      <td>86.0</td>\n",
              "      <td>217.081395</td>\n",
              "      <td>1455.441860</td>\n",
              "      <td>0.344182</td>\n",
              "      <td>0.575528</td>\n",
              "      <td>0.782883</td>\n",
              "      <td>0.624833</td>\n",
              "      <td>0.724872</td>\n",
              "      <td>0.872171</td>\n",
              "      <td>...</td>\n",
              "      <td>1.0</td>\n",
              "      <td>19.0</td>\n",
              "      <td>15.0</td>\n",
              "      <td>17.0</td>\n",
              "      <td>9.0</td>\n",
              "      <td>7.0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>9.0</td>\n",
              "      <td>13.123256</td>\n",
              "      <td>100.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>5 rows × 34 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "                                   code  #user_prompts  prompt_avg_chars  \\\n",
              "0  0031c86e-81f4-4eef-9e0e-28037abf9883           14.0        407.214286   \n",
              "1  0225686d-b825-4cac-8691-3a3a5343df2b           18.0        261.111111   \n",
              "2  041f950b-c013-409a-a642-cffff60b9d4b            9.0        663.000000   \n",
              "3  04f91058-d0f8-4324-83b2-19c671f433dc           20.0        114.100000   \n",
              "4  089eb66d-4c3a-4f58-b98f-a3774a2efb34           86.0        217.081395   \n",
              "\n",
              "   response_avg_chars       Q_0       Q_1       Q_2       Q_3       Q_4  \\\n",
              "0         2218.500000  0.153541  0.294879  0.522169  0.352243  0.460606   \n",
              "1         1653.111111  0.192013  0.795607  0.772184  0.882656  0.607114   \n",
              "2         1636.222222  0.258306  0.295923  0.624824  0.351872  0.643038   \n",
              "3         1749.900000  0.145965  0.117841  0.267346  0.316809  0.333889   \n",
              "4         1455.441860  0.344182  0.575528  0.782883  0.624833  0.724872   \n",
              "\n",
              "        Q_5  ...  Q_1_#prompts  Q_2_#prompts  Q_3_#prompts  Q_4_#prompts  \\\n",
              "0  0.553876  ...           1.0           2.0           3.0           6.0   \n",
              "1  0.987511  ...           1.0           5.0           1.0           4.0   \n",
              "2  0.454314  ...           0.0           2.0           3.0           1.0   \n",
              "3  0.309084  ...           0.0           6.0           3.0           6.0   \n",
              "4  0.872171  ...           1.0          19.0          15.0          17.0   \n",
              "\n",
              "   Q_5_#prompts  Q_6_#prompts  Q_7_#prompts  Q_8_#prompts  average_fk_grade  \\\n",
              "0           1.0           0.0           0.0           0.0         12.078571   \n",
              "1           2.0           1.0           3.0           1.0         12.872222   \n",
              "2           0.0           2.0           1.0           0.0         14.866667   \n",
              "3           1.0           1.0           2.0           1.0         11.030000   \n",
              "4           9.0           7.0           4.0           9.0         13.123256   \n",
              "\n",
              "   grade  \n",
              "0   48.0  \n",
              "1   99.0  \n",
              "2   90.0  \n",
              "3   97.0  \n",
              "4  100.0  \n",
              "\n",
              "[5 rows x 34 columns]"
            ]
          },
          "execution_count": 641,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# Calculate total number of code lines\n",
        "code_lines_df[\"total_#code_lines\"] = code_lines_df[\"Q_0_#code_lines\"].copy()\n",
        "for i in range(1, 9):\n",
        "    code_lines_df[\"total_#code_lines\"] = code_lines_df[\"total_#code_lines\"] + code_lines_df[f\"Q_{i}_#code_lines\"]\n",
        "\n",
        "merged_df = pd.merge(df, question_mapping_scores, on=\"code\", how=\"left\")\n",
        "merged_df = pd.merge(merged_df, code_lines_df, on=\"code\", how=\"left\")\n",
        "merged_df = pd.merge(merged_df, promptCountsDF, on=\"code\", how=\"left\")\n",
        "merged_df = pd.merge(merged_df, fk_grade_df, on=\"code\", how=\"left\")\n",
        "# merged_df = pd.merge(merged_df, engagement_scores_df, on=\"code\", how=\"left\")\n",
        "# merged_df = pd.merge(merged_df, message_variability_df, on=\"code\", how=\"left\")\n",
        "# merged_df = pd.merge(merged_df, sentiment_df, on=\"code\", how=\"left\")\n",
        "# merged_df[\"prompt_ratio\"] = prompt_ratio\n",
        "# merged_df[\"zero_prompt_counts\"] = zeros_count_per_row\n",
        "\n",
        "merged_df.drop(columns=[\"#error\", \"#no\", \"#next\", \"#thank\", \"#entropy\", \"#apologize\", \"#see\"], inplace=True)\n",
        "\n",
        "temp_df = pd.merge(merged_df, scores, on='code', how=\"left\")\n",
        "temp_df.dropna(inplace=True)\n",
        "temp_df.drop_duplicates(\"code\",inplace=True, keep=\"first\")\n",
        "temp_df.head()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FaryZpiytCmI"
      },
      "source": [
        "# 3. Model Training and Evaluation"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 642,
      "metadata": {},
      "outputs": [],
      "source": [
        "from skopt import BayesSearchCV\n",
        "from sklearn.model_selection import GridSearchCV\n",
        "from sklearn.feature_selection import SelectFromModel\n",
        "from xgboost import XGBRegressor, plot_importance\n",
        "from sklearn.ensemble import RandomForestRegressor\n",
        "from sklearn.tree import DecisionTreeRegressor"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 643,
      "metadata": {},
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "/var/folders/p4/w5cgd6hs1tsd7tddgjp26yn80000gn/T/ipykernel_56943/3383812256.py:2: FutureWarning: The default value of numeric_only in DataFrame.corr is deprecated. In a future version, it will default to False. Select only valid columns or specify the value of numeric_only to silence this warning.\n",
            "  correlations_with_target = temp_df.corr()['grade'].drop('grade')\n"
          ]
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 800x600 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "# Calculate correlations with the target variable\n",
        "correlations_with_target = temp_df.corr()['grade'].drop('grade')\n",
        "\n",
        "# Plot a bar plot to visualize correlations\n",
        "plt.figure(figsize=(8, 6))\n",
        "correlations_with_target.plot(kind='bar', color='skyblue')\n",
        "plt.title('Correlations with Grade')\n",
        "plt.xlabel('Features')\n",
        "plt.ylabel('Correlation')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 644,
      "metadata": {},
      "outputs": [],
      "source": [
        "# Create scatter plots for each feature with respect to the target variable\n",
        "features = temp_df.drop(['grade', 'code'], axis=1)\n",
        "\n",
        "for feature in features.columns:\n",
        "    plt.figure(figsize=(8, 6))\n",
        "    sns.scatterplot(x=feature, y='grade', data=temp_df, color='skyblue')\n",
        "    plt.title(f'{feature} vs Grade')\n",
        "    plt.xlabel(feature)\n",
        "    plt.ylabel('Grade')\n",
        "\n",
        "    # Save the plot as a PNG file\n",
        "    plt.savefig(f'plots/{feature}_scatter_plot.png', bbox_inches='tight')\n",
        "    plt.close()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 645,
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "(122, 32) (122,)\n"
          ]
        }
      ],
      "source": [
        "X = temp_df[temp_df.columns[1:-1]].to_numpy()\n",
        "y = temp_df[\"grade\"].to_numpy()\n",
        "print(X.shape, y.shape)\n",
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "## 4.1. Decision Tree"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 646,
      "metadata": {},
      "outputs": [
        {
          "data": {
            "text/html": [
              "<style>#sk-container-id-44 {color: black;background-color: white;}#sk-container-id-44 pre{padding: 0;}#sk-container-id-44 div.sk-toggleable {background-color: white;}#sk-container-id-44 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-44 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-44 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-44 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-44 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-44 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-44 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-44 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-44 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-44 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-44 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-44 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-44 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-44 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-44 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-44 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-44 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-44 div.sk-item {position: relative;z-index: 1;}#sk-container-id-44 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-44 div.sk-item::before, #sk-container-id-44 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-44 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-44 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-44 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-44 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-44 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-44 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-44 div.sk-label-container {text-align: center;}#sk-container-id-44 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-44 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-44\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>DecisionTreeRegressor()</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-44\" type=\"checkbox\" checked><label for=\"sk-estimator-id-44\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">DecisionTreeRegressor</label><div class=\"sk-toggleable__content\"><pre>DecisionTreeRegressor()</pre></div></div></div></div></div>"
            ],
            "text/plain": [
              "DecisionTreeRegressor()"
            ]
          },
          "execution_count": 646,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "regressor = DecisionTreeRegressor()\n",
        "regressor.fit(X_train, y_train)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 647,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LPC9Bn7tXJK9",
        "outputId": "bde79478-9c4b-4da0-8341-96b4f225c349"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "MSE Train: 0.0\n",
            "MSE TEST: 144.96\n",
            "R2 Train: 1.0\n",
            "R2 TEST: -0.29122366958355905\n"
          ]
        }
      ],
      "source": [
        "# Prediction\n",
        "y_train_pred = regressor.predict(X_train)\n",
        "y_test_pred = regressor.predict(X_test)\n",
        "\n",
        "# Calculation of Mean Squared Error (MSE)\n",
        "print(\"MSE Train:\", mean_squared_error(y_train,y_train_pred))\n",
        "print(\"MSE TEST:\", mean_squared_error(y_test,y_test_pred))\n",
        "\n",
        "print(\"R2 Train:\", r2_score(y_train,y_train_pred))\n",
        "print(\"R2 TEST:\", r2_score(y_test,y_test_pred))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 648,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "ofYUwWB_cCaY",
        "outputId": "fd4d85cd-71cb-40c1-a945-00459ba358b3"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 800x600 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 800x600 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 800x600 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAscAAAIjCAYAAADvI7a6AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjYuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8o6BhiAAAACXBIWXMAAA9hAAAPYQGoP6dpAABE5UlEQVR4nO3deXxN1/7/8ffJHEMSiQxCEDOlolIacyumjooOqVao8mvRKq5bOhjaktKrVdVS7TW06lbpcNVVt7m0VIWaSimpqiFFIqpJjBE56/dHt/N1ZBCRUV/Px+M8bs7aa+392dmr7tu2zzo2Y4wRAAAAALmUdgEAAABAWUE4BgAAACyEYwAAAMBCOAYAAAAshGMAAADAQjgGAAAALIRjAAAAwEI4BgAAACyEYwAAAMBCOAZw3ZkwYYJsNluB+tpsNk2YMKFY6+nUqZM6depUZveXn71796pr167y9fWVzWbT559/XiLHLS7z58+XzWbTgQMHSrsUAGUU4RhAsbkYRC6+3NzcVL16dfXv31+HDx8u7fLKnNq1azv9voKCgtS+fXt99tlnRbL/M2fOaMKECfrmm28KPCY2NlY//vijJk2apA8++ECRkZFFUsvlOnXq5HTueb2K+y8yhXH5HPf391fLli01fPhw/fTTT4Xeb2GuF4Br51baBQC4/r344osKDw/XuXPntGHDBs2fP1/r1q3Tzp075eXlVeTHe/755zVmzJgi329JiIiI0KhRoyRJR44c0TvvvKNevXpp1qxZevzxx69p32fOnNHEiRMlqUB3ns+ePauEhAQ999xzGjZs2DUd+0qee+45PfbYY473mzZt0owZM/Tss8+qcePGjvYbb7zxmo7zyCOP6MEHH5Snp+c17edyXbp0Ub9+/WSMUXp6urZv364FCxbo7bff1pQpUzRy5Mir3ufVXi8ARYNwDKDY9ejRw3HH8bHHHlPVqlU1ZcoULVu2TPfff3+RH8/NzU1ubuXzj7fq1avr4Ycfdrzv16+f6tWrp9dff/2aw/HVSk1NlST5+fkV2T5Pnz6tihUr5mjv0qWL03svLy/NmDFDXbp0yTcY5rW/vLi6usrV1bXA/QuqQYMGTtdNkl555RXdddddGjVqlBo1aqTbb7+9yI8LoOjxWAWAEte+fXtJ0r59+5za9+zZoz59+sjf319eXl6KjIzUsmXLnPpkZWVp4sSJql+/vry8vBQQEKB27dopPj7e0Se3Z44zMzM1YsQIBQYGqnLlyrr77rv122+/5aitf//+ql27do723PY5b9483XbbbQoKCpKnp6eaNGmiWbNmXdXv4kpCQkLUuHFj7d+/P99+x44d08CBAxUcHCwvLy81b95cCxYscGw/cOCAAgMDJUkTJ0684mMKEyZMUK1atSRJo0ePls1mc/q9bNu2TT169JCPj48qVaqkzp07a8OGDU77uPhYzZo1azRkyBAFBQWpRo0ahfgt/F9NNptNP/30kx566CFVqVJF7dq1kyTt2LFD/fv3V506deTl5aWQkBA9+uij+v3333Ot6dJnjmvXrq0777xT69atU6tWreTl5aU6dero/fffL3StkhQQEKCPPvpIbm5umjRpkqP9/PnzGjdunFq2bClfX19VrFhR7du319dff+3oc6XrVdDzBXD1yuetFQDl2sVgUqVKFUfbrl271LZtW1WvXl1jxoxRxYoV9fHHH6tnz5765JNPdO+990r6MyDFxcXpscceU6tWrZSRkaHNmzdr69atOe4+Xuqxxx7TwoUL9dBDD6lNmzZavXq17rjjjms6j1mzZumGG27Q3XffLTc3N33xxRcaMmSI7Ha7hg4dek37vigrK0tJSUkKCAjIs8/Zs2fVqVMn/fLLLxo2bJjCw8O1ZMkS9e/fX2lpaRo+fLgCAwM1a9YsPfHEE7r33nvVq1cvSXk/ptCrVy/5+flpxIgRiomJ0e23365KlSpJ+vNatW/fXj4+Pvr73/8ud3d3vfPOO+rUqZPWrFmj1q1bO+1ryJAhCgwM1Lhx43T69Olr/p3cd999ql+/viZPnixjjCQpPj5ev/76qwYMGKCQkBDt2rVLc+bM0a5du7Rhw4YrfkDzl19+UZ8+fTRw4EDFxsZq7ty56t+/v1q2bKkbbrih0LXWrFlTHTt21Ndff62MjAz5+PgoIyND7733nmJiYjRo0CCdPHlS//znP9WtWzd9//33ioiIuOL1utbzBZAPAwDFZN68eUaS+d///mdSU1NNUlKSWbp0qQkMDDSenp4mKSnJ0bdz586mWbNm5ty5c442u91u2rRpY+rXr+9oa968ubnjjjvyPe748ePNpX+8/fDDD0aSGTJkiFO/hx56yEgy48ePd7TFxsaaWrVqXXGfxhhz5syZHP26detm6tSp49TWsWNH07Fjx3xrNsaYWrVqma5du5rU1FSTmppqtm/fbh588EEjyTz55JN57m/69OlGklm4cKGj7fz58yYqKspUqlTJZGRkGGOMSU1NzXG++dm/f7+RZF599VWn9p49exoPDw+zb98+R9uRI0dM5cqVTYcOHRxtF69/u3btzIULFwp0zIuWLFliJJmvv/7a0XbxGsTExOTon9u1+Ne//mUkmbVr1+aoaf/+/Y62WrVq5eh37Ngx4+npaUaNGnXFWiWZoUOH5rl9+PDhRpLZvn27McaYCxcumMzMTKc+f/zxhwkODjaPPvqooy2/61XQ8wVw9XisAkCxi46OVmBgoMLCwtSnTx9VrFhRy5Ytc/wT+4kTJ7R69Wrdf//9OnnypI4fP67jx4/r999/V7du3bR3717H6hZ+fn7atWuX9u7dW+Djr1ixQpL01FNPObU//fTT13Re3t7ejp/T09N1/PhxdezYUb/++qvS09MLtc+vvvpKgYGBCgwMVPPmzbVkyRI98sgjmjJlSp5jVqxYoZCQEMXExDja3N3d9dRTT+nUqVNas2ZNoWrJTXZ2tr766iv17NlTderUcbRXq1ZNDz30kNatW6eMjAynMYMGDSrS53xze/b60mtx7tw5HT9+XLfccoskaevWrVfcZ5MmTRyP+0hSYGCgGjZsqF9//fWa6714x/3kyZOS/nzu2cPDQ5Jkt9t14sQJXbhwQZGRkQWqVbr28wWQNx6rAFDs3nrrLTVo0EDp6emaO3eu1q5d67RawC+//CJjjF544QW98MILue7j2LFjql69ul588UXdc889atCggZo2baru3bvrkUceyXcVg4MHD8rFxUV169Z1am/YsOE1ndd3332n8ePHKyEhQWfOnHHalp6eLl9f36veZ+vWrfXyyy/LZrOpQoUKaty48RU/EHfw4EHVr19fLi7O9zsurvJw8ODBq64jL6mpqTpz5kyuv7vGjRvLbrcrKSnJ6VGE8PDwIjt+Xvs7ceKEJk6cqI8++kjHjh1z2laQv6jUrFkzR1uVKlX0xx9/FL5Qy6lTpyRJlStXdrQtWLBA06ZN0549e5SVleVoL+jv6lrPF0DeCMcAil2rVq0cq1X07NlT7dq100MPPaTExERVqlRJdrtdkvS3v/1N3bp1y3Uf9erVkyR16NBB+/bt07///W999dVXeu+99/T6669r9uzZTkuBFVZez2pmZ2c7vd+3b586d+6sRo0a6bXXXlNYWJg8PDy0YsUKvf76645zulpVq1ZVdHR0ocaWVZfe5Syu/d1///1av369Ro8erYiICMe86t69e4GuRV53to31TPO12Llzp1xdXR3Bd+HCherfv7969uyp0aNHKygoSK6uroqLi8vxIdW8XOv5Asgb4RhAiboYAm699VbNnDlTY8aMcfzzvLu7e4GCob+/vwYMGKABAwbo1KlT6tChgyZMmJBnOK5Vq5bsdrv27dvndMczMTExR98qVaooLS0tR/vld1+/+OILZWZmatmyZU53HS9dcaCk1KpVSzt27JDdbne6e7xnzx7Hdinv4H81AgMDVaFChVx/d3v27JGLi4vCwsKu+ThX448//tCqVas0ceJEjRs3ztF+NY/eFJdDhw5pzZo1ioqKctw5Xrp0qerUqaNPP/3U6ZqMHz/eaWxe16ssny9wPeCZYwAlrlOnTmrVqpWmT5+uc+fOKSgoSJ06ddI777yjo0eP5uh/cb1dSTmWqqpUqZLq1aunzMzMPI/Xo0cPSdKMGTOc2qdPn56jb926dZWenq4dO3Y42o4ePZrjW+ou3mm89M5ienq65s2bl2cdxeX2229XcnKyFi9e7Gi7cOGC3nzzTVWqVEkdO3aUJFWoUEGScg3/BeXq6qquXbvq3//+t9NyaCkpKVq0aJHatWsnHx+fQu+/sDVJOe/y5nZ9S9KJEycUExOj7OxsPffcc4723OrduHGjEhISnMbndb3K6vkC1wvuHAMoFaNHj9Z9992n+fPn6/HHH9dbb72ldu3aqVmzZho0aJDq1KmjlJQUJSQk6LffftP27dsl/fnBqU6dOqlly5by9/fX5s2btXTp0ny/wS0iIkIxMTF6++23lZ6erjZt2mjVqlX65ZdfcvR98MEH9cwzz+jee+/VU089pTNnzmjWrFlq0KCB0wedunbtKg8PD9111136f//v/+nUqVN69913FRQUlGvAL06DBw/WO++8o/79+2vLli2qXbu2li5dqu+++07Tp0933LH09vZWkyZNtHjxYjVo0ED+/v5q2rSpmjZtelXHe/nllxUfH6927dppyJAhcnNz0zvvvKPMzExNnTq1OE4xXz4+PurQoYOmTp2qrKwsVa9eXV999dUV14YuSj///LMWLlwoY4wyMjK0fft2LVmyRKdOndJrr72m7t27O/reeeed+vTTT3Xvvffqjjvu0P79+zV79mw1adLE8XyylP/1Ku3zBa5rpbhSBoDr3MVlszZt2pRjW3Z2tqlbt66pW7euY5mvffv2mX79+pmQkBDj7u5uqlevbu68806zdOlSx7iXX37ZtGrVyvj5+Rlvb2/TqFEjM2nSJHP+/HlHn9yWXTt79qx56qmnTEBAgKlYsaK56667TFJSUq5LZX311VemadOmxsPDwzRs2NAsXLgw130uW7bM3HjjjcbLy8vUrl3bTJkyxcydOzfHUmFXs5TblZapy2t/KSkpZsCAAaZq1arGw8PDNGvWzMybNy/H2PXr15uWLVsaDw+PKy7rltdSbsYYs3XrVtOtWzdTqVIlU6FCBXPrrbea9evXO/XJ7/pfSX5LuaWmpubo/9tvv5l7773X+Pn5GV9fX3PfffeZI0eO5DjHvJZyy+33XtDrJsnxcnFxMX5+fqZFixZm+PDhZteuXTn62+12M3nyZFOrVi3j6elpWrRoYZYvX57rMoJ5Xa+Cni+Aq2czpgg+bQAAAABcB3jmGAAAALAQjgEAAAAL4RgAAACwEI4BAAAAC+EYAAAAsBCOAQAAAAtfAlIE7Ha7jhw5osqVKxfJ17MCAACgaBljdPLkSYWGhsrFJe/7w4TjInDkyBGFhYWVdhkAAAC4gqSkJNWoUSPP7YTjInDxq1mTkpLk4+NTytUAAADgchkZGQoLC3PktrwQjovAxUcpfHx8CMcAAABl2JUegeUDeQAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYyl04fuutt1S7dm15eXmpdevW+v777/Ptv2TJEjVq1EheXl5q1qyZVqxYkWffxx9/XDabTdOnTy/iqgEAAFAelKtwvHjxYo0cOVLjx4/X1q1b1bx5c3Xr1k3Hjh3Ltf/69esVExOjgQMHatu2berZs6d69uypnTt35uj72WefacOGDQoNDS3u0wAAAEAZVa7C8WuvvaZBgwZpwIABatKkiWbPnq0KFSpo7ty5ufZ/44031L17d40ePVqNGzfWSy+9pJtuukkzZ8506nf48GE9+eST+vDDD+Xu7l4SpwIAAIAyqNyE4/Pnz2vLli2Kjo52tLm4uCg6OloJCQm5jklISHDqL0ndunVz6m+32/XII49o9OjRuuGGGwpUS2ZmpjIyMpxeAAAAKP/KTTg+fvy4srOzFRwc7NQeHBys5OTkXMckJydfsf+UKVPk5uamp556qsC1xMXFydfX1/EKCwu7ijMBAABAWVVuwnFx2LJli9544w3Nnz9fNputwOPGjh2r9PR0xyspKakYqwQAAEBJKTfhuGrVqnJ1dVVKSopTe0pKikJCQnIdExISkm//b7/9VseOHVPNmjXl5uYmNzc3HTx4UKNGjVLt2rXzrMXT01M+Pj5OLwAAAJR/5SYce3h4qGXLllq1apWjzW63a9WqVYqKisp1TFRUlFN/SYqPj3f0f+SRR7Rjxw798MMPjldoaKhGjx6t//73v8V3MgAAACiT3Eq7gKsxcuRIxcbGKjIyUq1atdL06dN1+vRpDRgwQJLUr18/Va9eXXFxcZKk4cOHq2PHjpo2bZruuOMOffTRR9q8ebPmzJkjSQoICFBAQIDTMdzd3RUSEqKGDRuW7MkBAACg1JWrcPzAAw8oNTVV48aNU3JysiIiIrRy5UrHh+4OHTokF5f/uxnepk0bLVq0SM8//7yeffZZ1a9fX59//rmaNm1aWqcAAACAMsxmjDGlXUR5l5GRIV9fX6Wnp/P8MQAAQBlU0LxWbp45BgAAAIob4RgAAACwEI4BAAAAC+EYAAAAsBCOAQAAAAvhGAAAALAQjgEAAAAL4RgAAACwEI4BAAAAC+EYAAAAsBCOAQAAAAvhGAAAALAQjgEAAAAL4RgAAACwEI4BAAAAC+EYAAAAsBCOAQAAAAvhGAAAALAQjgEAAAAL4RgAAACwEI4BAAAAC+EYAAAAsBCOAQAAAAvhGAAAALAQjgEAAAAL4RgAAACwEI4BAAAAC+EYAAAAsBCOAQAAAAvhGAAAALAQjgEAAAAL4RgAAACwEI4BAAAAC+EYAAAAsBCOAQAAAAvhGAAAALAQjgEAAAAL4RgAAACwEI4BAAAAC+EYAAAAsBCOAQAAAAvhGAAAALAQjgEAAAAL4RgAAACwEI4BAAAAC+EYAAAAsBCOAQAAAAvhGAAAALAQjgEAAAAL4RgAAACwEI4BAAAAC+EYAAAAsBCOAQAAAAvhGAAAALAQjgEAAAAL4RgAAACwEI4BAAAAC+EYAAAAsBCOAQAAAAvhGAAAALAQjgEAAAAL4RgAAACwEI4BAAAAC+EYAAAAsBCOAQAAAAvhGAAAALAQjgEAAAAL4RgAAACwEI4BAAAAC+EYAAAAsBCOAQAAAAvhGAAAALAQjgEAAABLuQvHb731lmrXri0vLy+1bt1a33//fb79lyxZokaNGsnLy0vNmjXTihUrHNuysrL0zDPPqFmzZqpYsaJCQ0PVr18/HTlypLhPAwAAAGVQuQrHixcv1siRIzV+/Hht3bpVzZs3V7du3XTs2LFc+69fv14xMTEaOHCgtm3bpp49e6pnz57auXOnJOnMmTPaunWrXnjhBW3dulWffvqpEhMTdffdd5fkaQEAAKCMsBljTGkXUVCtW7fWzTffrJkzZ0qS7Ha7wsLC9OSTT2rMmDE5+j/wwAM6ffq0li9f7mi75ZZbFBERodmzZ+d6jE2bNqlVq1Y6ePCgatasWaC6MjIy5Ovrq/T0dPn4+BTizAAAAFCcCprXys2d4/Pnz2vLli2Kjo52tLm4uCg6OloJCQm5jklISHDqL0ndunXLs78kpaeny2azyc/PL88+mZmZysjIcHoBAACg/Cs34fj48ePKzs5WcHCwU3twcLCSk5NzHZOcnHxV/c+dO6dnnnlGMTEx+f6NIi4uTr6+vo5XWFjYVZ4NAAAAyqJyE46LW1ZWlu6//34ZYzRr1qx8+44dO1bp6emOV1JSUglVCQAAgOLkVtoFFFTVqlXl6uqqlJQUp/aUlBSFhITkOiYkJKRA/S8G44MHD2r16tVXfG7Y09NTnp6ehTgLAAAAlGXl5s6xh4eHWrZsqVWrVjna7Ha7Vq1apaioqFzHREVFOfWXpPj4eKf+F4Px3r179b///U8BAQHFcwIAAAAo88rNnWNJGjlypGJjYxUZGalWrVpp+vTpOn36tAYMGCBJ6tevn6pXr664uDhJ0vDhw9WxY0dNmzZNd9xxhz766CNt3rxZc+bMkfRnMO7Tp4+2bt2q5cuXKzs72/E8sr+/vzw8PErnRAEAAFAqylU4fuCBB5Samqpx48YpOTlZERERWrlypeNDd4cOHZKLy//dDG/Tpo0WLVqk559/Xs8++6zq16+vzz//XE2bNpUkHT58WMuWLZMkRUREOB3r66+/VqdOnUrkvAAAAFA2lKt1jssq1jkGAAAo2667dY4BAACA4kY4BgAAACyEYwAAAMBCOAYAAAAshGMAAADAQjgGAAAALIRjAAAAwEI4BgAAACyEYwAAAMBCOAYAAAAshGMAAADAQjgGAAAALIRjAAAAwEI4BgAAACyEYwAAAMBCOAYAAAAshGMAAADAQjgGAAAALIRjAAAAwEI4BgAAACyEYwAAAMBCOAYAAAAshGMAAADAQjgGAAAALIRjAAAAwEI4BgAAACyEYwAAAMBCOAYAAAAshGMAAADAQjgGAAAALIRjAAAAwEI4BgAAACyEYwAAAMBCOAYAAAAshGMAAADAQjgGAAAALIRjAAAAwEI4BgAAACyEYwAAAMBCOAYAAAAshGMAAADAQjgGAAAALIRjAAAAwEI4BgAAACyEYwAAAMBCOAYAAAAshGMAAADAUmThOC0trah2BQAAAJSKQoXjKVOmaPHixY73999/vwICAlS9enVt3769yIoDAAAASlKhwvHs2bMVFhYmSYqPj1d8fLy+/PJL9ejRQ6NHjy7SAgEAAICS4laYQcnJyY5wvHz5ct1///3q2rWrateurdatWxdpgQAAAEBJKdSd4ypVqigpKUmStHLlSkVHR0uSjDHKzs4uuuoAAACAElSoO8e9evXSQw89pPr16+v3339Xjx49JEnbtm1TvXr1irRAAAAAoKQUKhy//vrrql27tpKSkjR16lRVqlRJknT06FENGTKkSAsEAAAASorNGGNKu4jyLiMjQ76+vkpPT5ePj09plwMAAIDLFDSvFfjO8bJlywp88LvvvrvAfQEAAICyosDhuGfPngXqZ7PZ+FAeAAAAyqUCh2O73V6cdQAAAAClrsi+PhoAAAAo7wq1WoUknT59WmvWrNGhQ4d0/vx5p21PPfXUNRcGAAAAlLRCheNt27bp9ttv15kzZ3T69Gn5+/vr+PHjqlChgoKCggjHAAAAKJcK9VjFiBEjdNddd+mPP/6Qt7e3NmzYoIMHD6ply5b6xz/+UdQ1AgAAACWiUOH4hx9+0KhRo+Ti4iJXV1dlZmYqLCxMU6dO1bPPPlvUNQIAAAAlolDh2N3dXS4ufw4NCgrSoUOHJEm+vr5KSkoquuoAAACAElSoZ45btGihTZs2qX79+urYsaPGjRun48eP64MPPlDTpk2LukYAAACgRBTqzvHkyZNVrVo1SdKkSZNUpUoVPfHEE0pNTdWcOXOKtEAAAACgpNiMMaa0iyjvCvpd3QAAACgdBc1rfAkIAAAAYCnUM8fh4eGy2Wx5bv/1118LXRAAAABQWgoVjp9++mmn91lZWdq2bZtWrlyp0aNHF0VdAAAAQIkrVDgePnx4ru1vvfWWNm/efE0FAQAAAKWlSJ857tGjhz755JOi3CUAAABQYgp15zgvS5culb+/f1HuEpfJtht9v/+Ejp08p6DKXmoV7i9Xl7yf/0bp4DrhUswHXK4wc6Kw86go519J1pDfmLy2pZ/J0qPzv9eR9HMK9fXS3P6t5FvBXZJ09ny2Jq/4SQd+P6PaARX07O1N5O3hesVtV1tffvvKr7689nf+gl0fJBzQwRNnVMu/gh6Jqi0PN5cr7i+vbakZmbr37XU6cTpL/hXd9dmQdgr08cx3zKlzFzRi8TYd+uOsalbx1usPtFAlrz8jZHLaOd355lplnLsgHy83LX+yg0L8vCRJiUdO6vY31yrbSK42acWTHdQwtLK2/PqHes9Z7/hdfjK4jVrWqXLFeVRSCrWUW4sWLZw+kGeMUXJyslJTU/X2229r8ODBRVrkpd566y29+uqrSk5OVvPmzfXmm2+qVatWefZfsmSJXnjhBR04cED169fXlClTdPvttzvVPn78eL377rtKS0tT27ZtNWvWLNWvX7/ANZXUUm4rdx7VxC9+0tH0c462ar5eGn9XE3VvWq3Yjourw3XCpZgPuFxh5kRh51FRzr+SrCG/MZJy3XYh267UU+dz7KtWgLcaBFdW/E/Hcmzr0iRIkvLc9m6/m6+qvsDKHtrxW0au+/o55aQO/n421/rG9mic6/6aVvfRqt3HZL8kqbnYpEHtw7VyV3Ke+5OU6zabpNxCn4+Xm6pUdM91jIerTeezc466sYaP9qac0tkse45t3u4uubZfyYFX7rjqMVejoHmtUOF44sSJTu9dXFwUGBioTp06qVGjRldfbQEtXrxY/fr10+zZs9W6dWtNnz5dS5YsUWJiooKCgnL0X79+vTp06KC4uDjdeeedWrRokaZMmaKtW7c6vslvypQpiouL04IFCxQeHq4XXnhBP/74o3766Sd5eXkVqK6SCMcrdx7VEwu35pjUF/+KMuvhm/g/2jKA64RLMR9wucLMicLOo6KcfyVZQ35jSvqLGXILyHnVh6JRnAG5WMNxaWndurVuvvlmzZw5U5Jkt9sVFhamJ598UmPGjMnR/4EHHtDp06e1fPlyR9stt9yiiIgIzZ49W8YYhYaGatSoUfrb3/4mSUpPT1dwcLDmz5+vBx98sEB1OX7ZR44USzjOthtFv/aNktMzc91ukxTs66n/jezEP9WWIq4TLsV8wOUKMycKO4+Kcv6VZA1XGlMatr7QxfFYRFms73qz6NHWahFePI9YZGRkyDc0tOjCcUZGzn8myEtxBMTz58+rQoUKWrp0qXr27Oloj42NVVpamv7973/nGFOzZk2NHDnSaem58ePH6/PPP9f27dv166+/qm7dutq2bZsiIiIcfTp27KiIiAi98cYbudaSmZmpzMz/+w8jIyNDYWFhSpfE9+MBAACUPRmSfKUrhuMCfyDPz88v3y/+uFR2dnZBd1tgx48fV3Z2toKDg53ag4ODtWfPnlzHJCcn59o/OTnZsf1iW159chMXF5fj0RIAAACUfwUOx19//bXj5wMHDmjMmDHq37+/oqKiJEkJCQlasGCB4uLiir7KMmbs2LEaOXKk4/3FO8c6ckQqhrvmG3/9Xf3nbbpiv/kDblbrOgFFfnwUDNcJl2I+4HKFmROFnUdFOf9KsoaCjilJMa1qaNxdf35OqSzWdz3a/VL34tlxRoYUGnrFbgUOxx07dnT8/OKLL+q1115TTEyMo+3uu+9Ws2bNNGfOHMXGxl5ltVdWtWpVubq6KiUlxak9JSVFISEhuY4JCQnJt//F/01JSVG1atWc+lz6mMXlPD095enpmXNDxYp/vopY5A0V5Be4V8np53L9AIBNUoivlyJvCPvzY6woFVwnXIr5gMsVZk4Udh4V5fwryRquNKY0jO4VKVnPHJfF+q43nwxuUyxZSpJUwCcbCvUlIAkJCYqMjMzRHhkZqe+//74wu7wiDw8PtWzZUqtWrXK02e12rVq1ynH3+nJRUVFO/SUpPj7e0T88PFwhISFOfTIyMrRx48Y891kaXF1sjuVrLv8j7OL78Xc14UM9pYzrhEsxH3C5wsyJws6jopx/JVlDQcbktq24dGkS5LTecX71XSv+JPhTWVjvuFDhOCwsTO+++26O9vfee+/PxwuKyciRI/Xuu+9qwYIF2r17t5544gmdPn1aAwYMkCT169dPY8eOdfQfPny4Vq5cqWnTpmnPnj2aMGGCNm/erGHDhkmSbDabnn76ab388statmyZfvzxR/Xr10+hoaFOH/orC7o3raZZD9+kEF/n5eVCfL1YDqoM4TrhUswHXK4wc6Kw86go519J1pDfmNkP36TZuWyr5uulwEoeudZQK8DbsZ7x5bo0Ccp3W27rHOdVXzVfL91YI/dHK7s0CXKsP5xbfXmdU5cmQTlu7LvYpP/XITzf/eW1La8A7uPllucYD9fcR91Yw0fe7rnHyLzar6S41zkuqEIt5bZixQr17t1b9erVU+vWrSVJ33//vfbu3atPPvnE6Us2itrMmTMdXwISERGhGTNmOGro1KmTateurfnz5zv6L1myRM8//7zjS0CmTp2a65eAzJkzR2lpaWrXrp3efvttNWjQoMA1ldSXgEh801Z5wXXCpZgPuBzfkMc35PENeSX/DXnFvs5xUlKSZs2a5VgponHjxnr88ceL9c5xWVWS4RgAAABX77r8EpCyinAMAABQthU0rxV4tYodO3aoadOmcnFx0Y4dO/Lte+ONNxa8UgAAAKCMKHA4joiIUHJysoKCghQRESGbzabcbjrbbLZi+RIQAAAAoLgVOBzv379fgYGBjp8BAACA602Bw3GtWrVy/RkAAAC4XhRqIboFCxboP//5j+P93//+d/n5+alNmzY6ePBgkRUHAAAAlKRChePJkyfL2/vPxaITEhI0c+ZMTZ06VVWrVtWIESOKtEAAAACgpBT4sYpLJSUlqV69epKkzz//XH369NHgwYPVtm1bderUqSjrAwAAAEpMoe4cV6pUSb///rsk6auvvlKXLl0kSV5eXjp79mzRVQcAAACUoELdOe7SpYsee+wxtWjRQj///LPj65h37dql2rVrF2V9AAAAQIkp1J3jt956S1FRUUpNTdUnn3yigIAASdKWLVsUExNTpAUCAAAAJYWvjy4CfH00AABA2VbQvFaoO8eS9O233+rhhx9WmzZtdPjwYUnSBx98oHXr1hV2lwAAAECpKlQ4/uSTT9StWzd5e3tr69atyszMlCSlp6dr8uTJRVogAAAAUFIKFY5ffvllzZ49W++++67c3d0d7W3bttXWrVuLrDgAAACgJBUqHCcmJqpDhw452n19fZWWlnatNQEAAAClolDhOCQkRL/88kuO9nXr1qlOnTrXXBQAAABQGgoVjgcNGqThw4dr48aNstlsOnLkiD788EONGjVKTzzxRFHXCAAAAJSIQn0JyJgxY2S329W5c2edOXNGHTp0kKenp0aPHq3HHnusqGsEAAAASkSh7hzbbDY999xzOnHihHbu3KkNGzYoNTVVvr6+Cg8PL+oaAQAAgBJxVeE4MzNTY8eOVWRkpNq2basVK1aoSZMm2rVrlxo2bKg33nhDI0aMKK5aAQAAgGJ1VY9VjBs3Tu+8846io6O1fv163XfffRowYIA2bNigadOm6b777pOrq2tx1QoAAAAUq6sKx0uWLNH777+vu+++Wzt37tSNN96oCxcuaPv27bLZbMVVIwAAAFAiruqxit9++00tW7aUJDVt2lSenp4aMWIEwRgAAADXhasKx9nZ2fLw8HC8d3NzU6VKlYq8KAAAAKA0XNVjFcYY9e/fX56enpKkc+fO6fHHH1fFihWd+n366adFVyEAAABQQq4qHMfGxjq9f/jhh4u0GAAAAKA0XVU4njdvXnHVAQAAAJS6Qn0JCAAAAHA9IhwDAAAAFsIxAAAAYCEcAwAAABbCMQAAAGAhHAMAAAAWwjEAAABgIRwDAAAAFsIxAAAAYCEcAwAAABbCMQAAAGAhHAMAAAAWwjEAAABgIRwDAAAAFsIxAAAAYCEcAwAAABbCMQAAAGAhHAMAAAAWwjEAAABgIRwDAAAAFsIxAAAAYCEcAwAAABbCMQAAAGAhHAMAAAAWwjEAAABgIRwDAAAAFsIxAAAAYCEcAwAAABbCMQAAAGAhHAMAAAAWwjEAAABgIRwDAAAAFsIxAAAAYCEcAwAAABbCMQAAAGAhHAMAAAAWwjEAAABgIRwDAAAAFsIxAAAAYCEcAwAAABbCMQAAAGAhHAMAAAAWwjEAAABgIRwDAAAAFsIxAAAAYCk34fjEiRPq27evfHx85Ofnp4EDB+rUqVP5jjl37pyGDh2qgIAAVapUSb1791ZKSopj+/bt2xUTE6OwsDB5e3urcePGeuONN4r7VAAAAFBGlZtw3LdvX+3atUvx8fFavny51q5dq8GDB+c7ZsSIEfriiy+0ZMkSrVmzRkeOHFGvXr0c27ds2aKgoCAtXLhQu3bt0nPPPaexY8dq5syZxX06AAAAKINsxhhT2kVcye7du9WkSRNt2rRJkZGRkqSVK1fq9ttv12+//abQ0NAcY9LT0xUYGKhFixapT58+kqQ9e/aocePGSkhI0C233JLrsYYOHardu3dr9erVBa4vIyNDvr6+Sk9Pl4+PTyHOEAAAAMWpoHmtXNw5TkhIkJ+fnyMYS1J0dLRcXFy0cePGXMds2bJFWVlZio6OdrQ1atRINWvWVEJCQp7HSk9Pl7+/f771ZGZmKiMjw+kFAACA8q9chOPk5GQFBQU5tbm5ucnf31/Jycl5jvHw8JCfn59Te3BwcJ5j1q9fr8WLF1/xcY24uDj5+vo6XmFhYQU/GQAAAJRZpRqOx4wZI5vNlu9rz549JVLLzp07dc8992j8+PHq2rVrvn3Hjh2r9PR0xyspKalEagQAAEDxcivNg48aNUr9+/fPt0+dOnUUEhKiY8eOObVfuHBBJ06cUEhISK7jQkJCdP78eaWlpTndPU5JSckx5qefflLnzp01ePBgPf/881es29PTU56enlfsBwAAgPKlVMNxYGCgAgMDr9gvKipKaWlp2rJli1q2bClJWr16tex2u1q3bp3rmJYtW8rd3V2rVq1S7969JUmJiYk6dOiQoqKiHP127dql2267TbGxsZo0aVIRnBUAAADKq3KxWoUk9ejRQykpKZo9e7aysrI0YMAARUZGatGiRZKkw4cPq3Pnznr//ffVqlUrSdITTzyhFStWaP78+fLx8dGTTz4p6c9ni6U/H6W47bbb1K1bN7366quOY7m6uhYotF/EahUAAABlW0HzWqneOb4aH374oYYNG6bOnTvLxcVFvXv31owZMxzbs7KylJiYqDNnzjjaXn/9dUffzMxMdevWTW+//bZj+9KlS5WamqqFCxdq4cKFjvZatWrpwIEDJXJeAAAAKDvKzZ3jsow7xwAAAGXbdbXOMQAAAFASCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYyk04PnHihPr27SsfHx/5+flp4MCBOnXqVL5jzp07p6FDhyogIECVKlVS7969lZKSkmvf33//XTVq1JDNZlNaWloxnAEAAADKunITjvv27atdu3YpPj5ey5cv19q1azV48OB8x4wYMUJffPGFlixZojVr1ujIkSPq1atXrn0HDhyoG2+8sThKBwAAQDlhM8aY0i7iSnbv3q0mTZpo06ZNioyMlCStXLlSt99+u3777TeFhobmGJOenq7AwEAtWrRIffr0kSTt2bNHjRs3VkJCgm655RZH31mzZmnx4sUaN26cOnfurD/++EN+fn4Fri8jI0O+vr5KT0+Xj4/PtZ0sAAAAilxB81q5uHOckJAgPz8/RzCWpOjoaLm4uGjjxo25jtmyZYuysrIUHR3taGvUqJFq1qyphIQER9tPP/2kF198Ue+//75cXAr268jMzFRGRobTCwAAAOVfuQjHycnJCgoKcmpzc3OTv7+/kpOT8xzj4eGR4w5wcHCwY0xmZqZiYmL06quvqmbNmgWuJy4uTr6+vo5XWFjY1Z0QAAAAyqRSDcdjxoyRzWbL97Vnz55iO/7YsWPVuHFjPfzww1c9Lj093fFKSkoqpgoBAABQktxK8+CjRo1S//798+1Tp04dhYSE6NixY07tFy5c0IkTJxQSEpLruJCQEJ0/f15paWlOd49TUlIcY1avXq0ff/xRS5culSRdfPy6atWqeu655zRx4sRc9+3p6SlPT8+CnCIAAADKkVINx4GBgQoMDLxiv6ioKKWlpWnLli1q2bKlpD+Drd1uV+vWrXMd07JlS7m7u2vVqlXq3bu3JCkxMVGHDh1SVFSUJOmTTz7R2bNnHWM2bdqkRx99VN9++63q1q17racHAACAcqZUw3FBNW7cWN27d9egQYM0e/ZsZWVladiwYXrwwQcdK1UcPnxYnTt31vvvv69WrVrJ19dXAwcO1MiRI+Xv7y8fHx89+eSTioqKcqxUcXkAPn78uON4V7NaBQAAAK4P5SIcS9KHH36oYcOGqXPnznJxcVHv3r01Y8YMx/asrCwlJibqzJkzjrbXX3/d0TczM1PdunXT22+/XRrlAwAAoBwoF+scl3WscwwAAFC2XVfrHAMAAAAlgXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAhXAMAAAAWAjHAAAAgIVwDAAAAFgIxwAAAICFcAwAAABYCMcAAACAxa20C7geGGMkSRkZGaVcCQAAAHJzMaddzG15IRwXgZMnT0qSwsLCSrkSAAAA5OfkyZPy9fXNc7vNXCk+44rsdruOHDmiypUry2azlXY5JSIjI0NhYWFKSkqSj49PaZeDUsZ8wOWYE7gccwKXKo35YIzRyZMnFRoaKheXvJ8s5s5xEXBxcVGNGjVKu4xS4ePjwx9ycGA+4HLMCVyOOYFLlfR8yO+O8UV8IA8AAACwEI4BAAAAC+EYheLp6anx48fL09OztEtBGcB8wOWYE7gccwKXKsvzgQ/kAQAAABbuHAMAAAAWwjEAAABgIRwDAAAAFsIxAAAAYCEcI09xcXG6+eabVblyZQUFBalnz55KTEx06nPu3DkNHTpUAQEBqlSpknr37q2UlJRSqhgl6ZVXXpHNZtPTTz/taGM+/PUcPnxYDz/8sAICAuTt7a1mzZpp8+bNju3GGI0bN07VqlWTt7e3oqOjtXfv3lKsGMUpOztbL7zwgsLDw+Xt7a26devqpZde0qWf/WdOXN/Wrl2ru+66S6GhobLZbPr888+dthfk+p84cUJ9+/aVj4+P/Pz8NHDgQJ06darEzoFwjDytWbNGQ4cO1YYNGxQfH6+srCx17dpVp0+fdvQZMWKEvvjiCy1ZskRr1qzRkSNH1KtXr1KsGiVh06ZNeuedd3TjjTc6tTMf/lr++OMPtW3bVu7u7vryyy/1008/adq0aapSpYqjz9SpUzVjxgzNnj1bGzduVMWKFdWtWzedO3euFCtHcZkyZYpmzZqlmTNnavfu3ZoyZYqmTp2qN99809GHOXF9O336tJo3b6633nor1+0Fuf59+/bVrl27FB8fr+XLl2vt2rUaPHhwSZ2CZIACOnbsmJFk1qxZY4wxJi0tzbi7u5slS5Y4+uzevdtIMgkJCaVVJorZyZMnTf369U18fLzp2LGjGT58uDGG+fBX9Mwzz5h27drlud1ut5uQkBDz6quvOtrS0tKMp6en+de//lUSJaKE3XHHHebRRx91auvVq5fp27evMYY58VcjyXz22WeO9wW5/j/99JORZDZt2uTo8+WXXxqbzWYOHz5cInVz5xgFlp6eLkny9/eXJG3ZskVZWVmKjo529GnUqJFq1qyphISEUqkRxW/o0KG64447nK67xHz4K1q2bJkiIyN13333KSgoSC1atNC7777r2L5//34lJyc7zQlfX1+1bt2aOXGdatOmjVatWqWff/5ZkrR9+3atW7dOPXr0kMSc+KsryPVPSEiQn5+fIiMjHX2io6Pl4uKijRs3lkidbiVyFJR7drtdTz/9tNq2baumTZtKkpKTk+Xh4SE/Pz+nvsHBwUpOTi6FKlHcPvroI23dulWbNm3KsY358Nfz66+/atasWRo5cqSeffZZbdq0SU899ZQ8PDwUGxvruO7BwcFO45gT168xY8YoIyNDjRo1kqurq7KzszVp0iT17dtXkpgTf3EFuf7JyckKCgpy2u7m5iZ/f/8SmyOEYxTI0KFDtXPnTq1bt660S0EpSUpK0vDhwxUfHy8vL6/SLgdlgN1uV2RkpCZPnixJatGihXbu3KnZs2crNja2lKtDafj444/14YcfatGiRbrhhhv0ww8/6Omnn1ZoaChzAuUGj1XgioYNG6bly5fr66+/Vo0aNRztISEhOn/+vNLS0pz6p6SkKCQkpISrRHHbsmWLjh07pptuuklubm5yc3PTmjVrNGPGDLm5uSk4OJj58BdTrVo1NWnSxKmtcePGOnTokCQ5rvvlK5YwJ65fo0eP1pgxY/Tggw+qWbNmeuSRRzRixAjFxcVJYk781RXk+oeEhOjYsWNO2y9cuKATJ06U2BwhHCNPxhgNGzZMn332mVavXq3w8HCn7S1btpS7u7tWrVrlaEtMTNShQ4cUFRVV0uWimHXu3Fk//vijfvjhB8crMjJSffv2dfzMfPhradu2bY7lHX/++WfVqlVLkhQeHq6QkBCnOZGRkaGNGzcyJ65TZ86ckYuLc7RwdXWV3W6XxJz4qyvI9Y+KilJaWpq2bNni6LN69WrZ7Xa1bt26ZAotkY/9oVx64oknjK+vr/nmm2/M0aNHHa8zZ844+jz++OOmZs2aZvXq1Wbz5s0mKirKREVFlWLVKEmXrlZhDPPhr+b77783bm5uZtKkSWbv3r3mww8/NBUqVDALFy509HnllVeMn5+f+fe//2127Nhh7rnnHhMeHm7Onj1bipWjuMTGxprq1aub5cuXm/3795tPP/3UVK1a1fz973939GFOXN9Onjxptm3bZrZt22Ykmddee81s27bNHDx40BhTsOvfvXt306JFC7Nx40azbt06U79+fRMTE1Ni50A4Rp4k5fqaN2+eo8/Zs2fNkCFDTJUqVUyFChXMvffea44ePVp6RaNEXR6OmQ9/PV988YVp2rSp8fT0NI0aNTJz5sxx2m63280LL7xggoODjaenp+ncubNJTEwspWpR3DIyMszw4cNNzZo1jZeXl6lTp4557rnnTGZmpqMPc+L69vXXX+eaHWJjY40xBbv+v//+u4mJiTGVKlUyPj4+ZsCAAebkyZMldg42Yy752hoAAADgL4xnjgEAAAAL4RgAAACwEI4BAAAAC+EYAAAAsBCOAQAAAAvhGAAAALAQjgEAAAAL4RgAAACwEI4BoIzr37+/evbs6XjfqVMnPf300yVexzfffCObzaa0tLQSPzYAlBTCMQAUQv/+/WWz2WSz2eTh4aF69erpxRdf1IULF4r92J9++qleeumlAvUt6UC7fft23X333QoKCpKXl5dq166tBx54QMeOHSuR4wPAtSIcA0Ahde/eXUePHtXevXs1atQoTZgwQa+++mqufc+fP19kx/X391flypWLbH9FJTU1VZ07d5a/v7/++9//avfu3Zo3b55CQ0N1+vTpYjtuVlZWse0bwF8P4RgACsnT01MhISGqVauWnnjiCUVHR2vZsmWS/u9RiEmTJik0NFQNGzaUJCUlJen++++Xn5+f/P39dc899+jAgQOOfWZnZ2vkyJHy8/NTQECA/v73v8sY43Tcyx+ryMzM1DPPPKOwsDB5enqqXr16+uc//6kDBw7o1ltvlSRVqVJFNptN/fv3lyTZ7XbFxcUpPDxc3t7eat68uZYuXep0nBUrVqhBgwby9vbWrbfe6lRnbr777julp6frvffeU4sWLRQeHq5bb71Vr7/+usLDwx39du3apTvvvFM+Pj6qXLmy2rdvr3379jnqevHFF1WjRg15enoqIiJCK1eudIw9cOCAbDabFi9erI4dO8rLy0sffvihJOm9995T48aN5eXlpUaNGuntt992jDt//ryGDRumatWqycvLS7Vq1VJcXFy+5wPgr8mttAsAgOuFt7e3fv/9d8f7VatWycfHR/Hx8ZL+vMPZrVs3RUVF6dtvv5Wbm5tefvllde/eXTt27JCHh4emTZum+fPna+7cuWrcuLGmTZumzz77TLfddluex+3Xr58SEhI0Y8YMNW/eXPv379fx48cVFhamTz75RL1791ZiYqJ8fHzk7e0tSYqLi9PChQs1e/Zs1a9fX2vXrtXDDz+swMBAdezYUUlJSerVq5eGDh2qwYMHa/PmzRo1alS+5x8SEqILFy7os88+U58+fWSz2XL0OXz4sDp06KBOnTpp9erV8vHx0Xfffed4HOWNN97QtGnT9M4776hFixaaO3eu7r77bu3atUv169d37GfMmDGaNm2aWrRo4QjI48aN08yZM9WiRQtt27ZNgwYNUsWKFRUbG6sZM2Zo2bJl+vjjj1WzZk0lJSUpKSmp4BcXwF+HAQBctdjYWHPPPfcYY4yx2+0mPj7eeHp6mr/97W+O7cHBwSYzM9Mx5oMPPjANGzY0drvd0ZaZmWm8vb3Nf//7X2OMMdWqVTNTp051bM/KyjI1atRwHMsYYzp27GiGDx9ujDEmMTHRSDLx8fG51vn1118bSeaPP/5wtJ07d85UqFDBrF+/3qnvwIEDTUxMjDHGmLFjx5omTZo4bX/mmWdy7Otyzz77rHFzczP+/v6me/fuZurUqSY5OdmxfezYsSY8PNycP38+1/GhoaFm0qRJTm0333yzGTJkiDHGmP379xtJZvr06U596tataxYtWuTU9tJLL5moqChjjDFPPvmkue2225x+9wCQG+4cA0AhLV++XJUqVVJWVpbsdrseeughTZgwwbG9WbNm8vDwcLzfvn27fvnllxzPC587d0779u1Tenq6jh49qtatWzu2ubm5KTIyMsejFRf98MMPcnV1VceOHQtc9y+//KIzZ86oS5cuTu3nz59XixYtJEm7d+92qkOSoqKirrjvSZMmaeTIkVq9erU2btyo2bNna/LkyVq7dq2aNWumH374Qe3bt5e7u3uOsRkZGTpy5Ijatm3r1N62bVtt377dqS0yMtLx8+nTp7Vv3z4NHDhQgwYNcrRfuHBBvr6+kv58zKVLly5q2LChunfvrjvvvFNdu3a94vkA+OshHANAId16662aNWuWPDw8FBoaKjc35z9SK1as6PT+1KlTatmypeMZ2UsFBgYWqoaLj0lcjVOnTkmS/vOf/6h69epO2zw9PQtVx6UCAgJ033336b777tPkyZPVokUL/eMf/9CCBQsKVW9uLv3dXjyfd999N0egd3V1lSTddNNN2r9/v7788kv973//0/3336/o6Ogcz1kDAOEYAAqpYsWKqlevXoH733TTTVq8eLGCgoLk4+OTa59q1app48aN6tChg6Q/735u2bJFN910U679mzVrJrvdrjVr1ig6OjrH9ot3rrOzsx1tTZo0kaenpw4dOpTnHefGjRs7Plx40YYNG658krkcv27duo7VKm688UYtWLBAWVlZOe4e+/j4KDQ0VN99951TXd99951atWqV5zGCg4MVGhqqX3/9VX379s2zn4+Pjx544AE98MAD6tOnj7p3764TJ07I39//qs8LwPWL1SoAoIT07dtXVatW1T333KNvv/1W+/fv1zfffKOnnnpKv/32myRp+PDheuWVV/T5559rz549GjJkSL5rFNeuXVuxsbF69NFH9fnnnzv2+fHHH0uSatWqJZvNpuXLlys1NVWnTp1S5cqV9be//U0jRozQggULtG/fPm3dulVvvvmmFixYIEl6/PHHtXfvXo0ePVqJiYlatGiR5s+fn+/5LV++XA8//LCWL1+un3/+WYmJifrHP/6hFStW6J577pEkDRs2TBkZGXrwwQe1efNm7d27Vx988IESExMlSaNHj9aUKVO0ePFiJSYmasyYMfrhhx80fPjwfI89ceJExcXFacaMGfr555/1448/at68eXrttdckSa+99pr+9a9/ac+ePfr555+1ZMkShYSEyM/P70qXDcBfTWk/9AwA5dGlH8i7mu1Hjx41/fr1M1WrVjWenp6mTp06ZtCgQSY9Pd0Y8+cH8IYPH258fHyMn5+fGTlypOnXr1+eH8gzxpizZ8+aESNGmGrVqhkPDw9Tr149M3fuXMf2F1980YSEhBibzWZiY2ONMX9+iHD69OmmYcOGxt3d3QQGBppu3bqZNWvWOMZ98cUXpl69esbT09O0b9/ezJ07N98P5O3bt88MGjTINGjQwHh7exs/Pz9z8803m3nz5jn12759u+nataupUKGCqVy5smnfvr3Zt2+fMcaY7OxsM2HCBFO9enXj7u5umjdvbr788kvH2IsfyNu2bVuO43/44YcmIiLCeHh4mCpVqpgOHTqYTz/91BhjzJw5c0xERISpWLGi8fHxMZ07dzZbt27N9TwA/LXZjMnjUx4AAADAXwyPVQAAAAAWwjEAAABgIRwDAAAAFsIxAAAAYCEcAwAAABbCMQAAAGAhHAMAAAAWwjEAAABgIRwDAAAAFsIxAAAAYCEcAwAAAJb/D1so6T0xCb90AAAAAElFTkSuQmCC",
            "text/plain": [
              "<Figure size 800x600 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "\n",
        "# Scatter plot for Test Data\n",
        "plt.figure(figsize=(8, 6))\n",
        "plt.scatter(y_test, y_test_pred)\n",
        "plt.xlabel('Actual Scores')\n",
        "plt.ylabel('Predicted Scores')\n",
        "plt.title('Actual vs Predicted Scores for Test Data')\n",
        "plt.show()\n",
        "\n",
        "# Scatter plot for Train Data\n",
        "plt.figure(figsize=(8, 6))\n",
        "plt.scatter(y_train, y_train_pred)\n",
        "plt.xlabel('Actual Scores')\n",
        "plt.ylabel('Predicted Scores')\n",
        "plt.title('Actual vs Predicted Scores for Train Data')\n",
        "plt.show()\n",
        "\n",
        "# Residual Plot for Test Data\n",
        "plt.figure(figsize=(8, 6))\n",
        "residuals_test = y_test - y_test_pred\n",
        "plt.scatter(y_test_pred, residuals_test)\n",
        "plt.xlabel('Predicted Scores')\n",
        "plt.ylabel('Residuals')\n",
        "plt.title('Residual Plot for Test Data')\n",
        "plt.axhline(y=0, color='red', linestyle='-')\n",
        "plt.show()\n",
        "\n",
        "# Residual Plot for Train Data\n",
        "plt.figure(figsize=(8, 6))\n",
        "residuals_train = y_train - y_train_pred\n",
        "plt.scatter(y_train_pred, residuals_train)\n",
        "plt.xlabel('Predicted Scores')\n",
        "plt.ylabel('Residuals')\n",
        "plt.title('Residual Plot for Train Data')\n",
        "plt.axhline(y=0, color='red', linestyle='-')\n",
        "plt.show()\n",
        "\n",
        "#Heatmap of Feature Correlations (assuming X_train is a DataFrame)\n",
        "#Heatmap için X_traini DataFrame dönüştürmek lazım yaparız\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "## 4.2. Random Forest"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 649,
      "metadata": {},
      "outputs": [
        {
          "data": {
            "text/html": [
              "<style>#sk-container-id-45 {color: black;background-color: white;}#sk-container-id-45 pre{padding: 0;}#sk-container-id-45 div.sk-toggleable {background-color: white;}#sk-container-id-45 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-45 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-45 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-45 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-45 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-45 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-45 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-45 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-45 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-45 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-45 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-45 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-45 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-45 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-45 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-45 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-45 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-45 div.sk-item {position: relative;z-index: 1;}#sk-container-id-45 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-45 div.sk-item::before, #sk-container-id-45 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-45 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-45 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-45 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-45 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-45 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-45 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-45 div.sk-label-container {text-align: center;}#sk-container-id-45 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-45 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-45\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>RandomForestRegressor(random_state=42)</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-45\" type=\"checkbox\" checked><label for=\"sk-estimator-id-45\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">RandomForestRegressor</label><div class=\"sk-toggleable__content\"><pre>RandomForestRegressor(random_state=42)</pre></div></div></div></div></div>"
            ],
            "text/plain": [
              "RandomForestRegressor(random_state=42)"
            ]
          },
          "execution_count": 649,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "rf_regressor = RandomForestRegressor(random_state=42)\n",
        "rf_regressor.fit(X_train, y_train)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 650,
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "MSE Train: 29.05831134020618\n",
            "MSE TEST: 113.79342000000001\n",
            "R2 Train: 0.8221866938688442\n",
            "R2 TEST: -0.01360897728244459\n"
          ]
        }
      ],
      "source": [
        "# Prediction\n",
        "y_train_pred = rf_regressor.predict(X_train)\n",
        "y_test_pred = rf_regressor.predict(X_test)\n",
        "\n",
        "# Calculation of Mean Squared Error (MSE)\n",
        "print(\"MSE Train:\", mean_squared_error(y_train,y_train_pred))\n",
        "print(\"MSE TEST:\", mean_squared_error(y_test,y_test_pred))\n",
        "\n",
        "print(\"R2 Train:\", r2_score(y_train,y_train_pred))\n",
        "print(\"R2 TEST:\", r2_score(y_test,y_test_pred))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "### Tune the Model"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "#### 1. Bayesian Search"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 651,
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=1, max_features=0.7117798785370942, min_samples_leaf=4, min_samples_split=9, n_estimators=142; total time=   0.2s\n",
            "[CV] END max_depth=1, max_features=0.7117798785370942, min_samples_leaf=4, min_samples_split=9, n_estimators=142; total time=   0.2s\n",
            "[CV] END max_depth=1, max_features=0.7117798785370942, min_samples_leaf=4, min_samples_split=9, n_estimators=142; total time=   0.2s\n",
            "[CV] END max_depth=1, max_features=0.7117798785370942, min_samples_leaf=4, min_samples_split=9, n_estimators=142; total time=   0.2s\n",
            "[CV] END max_depth=1, max_features=0.7117798785370942, min_samples_leaf=4, min_samples_split=9, n_estimators=142; total time=   0.2s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=6, max_features=0.13099903413168362, min_samples_leaf=7, min_samples_split=7, n_estimators=154; total time=   0.3s\n",
            "[CV] END max_depth=6, max_features=0.13099903413168362, min_samples_leaf=7, min_samples_split=7, n_estimators=154; total time=   0.3s\n",
            "[CV] END max_depth=6, max_features=0.13099903413168362, min_samples_leaf=7, min_samples_split=7, n_estimators=154; total time=   0.3s\n",
            "[CV] END max_depth=6, max_features=0.13099903413168362, min_samples_leaf=7, min_samples_split=7, n_estimators=154; total time=   0.1s\n",
            "[CV] END max_depth=6, max_features=0.13099903413168362, min_samples_leaf=7, min_samples_split=7, n_estimators=154; total time=   0.1s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=3, max_features=0.7773542320027641, min_samples_leaf=7, min_samples_split=11, n_estimators=89; total time=   0.1s\n",
            "[CV] END max_depth=3, max_features=0.7773542320027641, min_samples_leaf=7, min_samples_split=11, n_estimators=89; total time=   0.1s\n",
            "[CV] END max_depth=3, max_features=0.7773542320027641, min_samples_leaf=7, min_samples_split=11, n_estimators=89; total time=   0.1s\n",
            "[CV] END max_depth=3, max_features=0.7773542320027641, min_samples_leaf=7, min_samples_split=11, n_estimators=89; total time=   0.1s\n",
            "[CV] END max_depth=3, max_features=0.7773542320027641, min_samples_leaf=7, min_samples_split=11, n_estimators=89; total time=   0.2s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=8, max_features=0.32677209595677636, min_samples_leaf=3, min_samples_split=7, n_estimators=247; total time=   0.4s\n",
            "[CV] END max_depth=8, max_features=0.32677209595677636, min_samples_leaf=3, min_samples_split=7, n_estimators=247; total time=   0.4s\n",
            "[CV] END max_depth=8, max_features=0.32677209595677636, min_samples_leaf=3, min_samples_split=7, n_estimators=247; total time=   0.4s\n",
            "[CV] END max_depth=8, max_features=0.32677209595677636, min_samples_leaf=3, min_samples_split=7, n_estimators=247; total time=   0.4s\n",
            "[CV] END max_depth=8, max_features=0.32677209595677636, min_samples_leaf=3, min_samples_split=7, n_estimators=247; total time=   0.5s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=19, max_features=0.7898918835683693, min_samples_leaf=1, min_samples_split=18, n_estimators=66; total time=   0.2s\n",
            "[CV] END max_depth=19, max_features=0.7898918835683693, min_samples_leaf=1, min_samples_split=18, n_estimators=66; total time=   0.2s\n",
            "[CV] END max_depth=19, max_features=0.7898918835683693, min_samples_leaf=1, min_samples_split=18, n_estimators=66; total time=   0.2s\n",
            "[CV] END max_depth=19, max_features=0.7898918835683693, min_samples_leaf=1, min_samples_split=18, n_estimators=66; total time=   0.2s\n",
            "[CV] END max_depth=19, max_features=0.7898918835683693, min_samples_leaf=1, min_samples_split=18, n_estimators=66; total time=   0.2s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=14, max_features=0.9429163766748473, min_samples_leaf=2, min_samples_split=5, n_estimators=192; total time=   0.4s\n",
            "[CV] END max_depth=14, max_features=0.9429163766748473, min_samples_leaf=2, min_samples_split=5, n_estimators=192; total time=   0.4s\n",
            "[CV] END max_depth=14, max_features=0.9429163766748473, min_samples_leaf=2, min_samples_split=5, n_estimators=192; total time=   0.4s\n",
            "[CV] END max_depth=14, max_features=0.9429163766748473, min_samples_leaf=2, min_samples_split=5, n_estimators=192; total time=   0.4s\n",
            "[CV] END max_depth=14, max_features=0.9429163766748473, min_samples_leaf=2, min_samples_split=5, n_estimators=192; total time=   0.4s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=10, max_features=0.8313199612147388, min_samples_leaf=10, min_samples_split=23, n_estimators=87; total time=   0.1s\n",
            "[CV] END max_depth=10, max_features=0.8313199612147388, min_samples_leaf=10, min_samples_split=23, n_estimators=87; total time=   0.1s\n",
            "[CV] END max_depth=10, max_features=0.8313199612147388, min_samples_leaf=10, min_samples_split=23, n_estimators=87; total time=   0.2s\n",
            "[CV] END max_depth=10, max_features=0.8313199612147388, min_samples_leaf=10, min_samples_split=23, n_estimators=87; total time=   0.2s\n",
            "[CV] END max_depth=10, max_features=0.8313199612147388, min_samples_leaf=10, min_samples_split=23, n_estimators=87; total time=   0.2s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=2, max_features=0.9350841473488251, min_samples_leaf=7, min_samples_split=24, n_estimators=90; total time=   0.2s\n",
            "[CV] END max_depth=2, max_features=0.9350841473488251, min_samples_leaf=7, min_samples_split=24, n_estimators=90; total time=   0.2s\n",
            "[CV] END max_depth=2, max_features=0.9350841473488251, min_samples_leaf=7, min_samples_split=24, n_estimators=90; total time=   0.2s\n",
            "[CV] END max_depth=2, max_features=0.9350841473488251, min_samples_leaf=7, min_samples_split=24, n_estimators=90; total time=   0.2s\n",
            "[CV] END max_depth=2, max_features=0.9350841473488251, min_samples_leaf=7, min_samples_split=24, n_estimators=90; total time=   0.2s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=19, max_features=0.20125056930790036, min_samples_leaf=3, min_samples_split=20, n_estimators=185; total time=   0.3s\n",
            "[CV] END max_depth=19, max_features=0.20125056930790036, min_samples_leaf=3, min_samples_split=20, n_estimators=185; total time=   0.3s\n",
            "[CV] END max_depth=19, max_features=0.20125056930790036, min_samples_leaf=3, min_samples_split=20, n_estimators=185; total time=   0.3s\n",
            "[CV] END max_depth=19, max_features=0.20125056930790036, min_samples_leaf=3, min_samples_split=20, n_estimators=185; total time=   0.3s\n",
            "[CV] END max_depth=19, max_features=0.20125056930790036, min_samples_leaf=3, min_samples_split=20, n_estimators=185; total time=   0.3s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=3, max_features=0.7727334535385882, min_samples_leaf=9, min_samples_split=17, n_estimators=68; total time=   0.1s\n",
            "[CV] END max_depth=3, max_features=0.7727334535385882, min_samples_leaf=9, min_samples_split=17, n_estimators=68; total time=   0.1s\n",
            "[CV] END max_depth=3, max_features=0.7727334535385882, min_samples_leaf=9, min_samples_split=17, n_estimators=68; total time=   0.1s\n",
            "[CV] END max_depth=3, max_features=0.7727334535385882, min_samples_leaf=9, min_samples_split=17, n_estimators=68; total time=   0.1s\n",
            "[CV] END max_depth=3, max_features=0.7727334535385882, min_samples_leaf=9, min_samples_split=17, n_estimators=68; total time=   0.1s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=19, max_features=0.1, min_samples_leaf=4, min_samples_split=25, n_estimators=184; total time=   0.2s\n",
            "[CV] END max_depth=19, max_features=0.1, min_samples_leaf=4, min_samples_split=25, n_estimators=184; total time=   0.2s\n",
            "[CV] END max_depth=19, max_features=0.1, min_samples_leaf=4, min_samples_split=25, n_estimators=184; total time=   0.2s\n",
            "[CV] END max_depth=19, max_features=0.1, min_samples_leaf=4, min_samples_split=25, n_estimators=184; total time=   0.2s\n",
            "[CV] END max_depth=19, max_features=0.1, min_samples_leaf=4, min_samples_split=25, n_estimators=184; total time=   0.2s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=5, max_features=0.5000637862320251, min_samples_leaf=7, min_samples_split=25, n_estimators=165; total time=   0.4s\n",
            "[CV] END max_depth=5, max_features=0.5000637862320251, min_samples_leaf=7, min_samples_split=25, n_estimators=165; total time=   0.4s\n",
            "[CV] END max_depth=5, max_features=0.5000637862320251, min_samples_leaf=7, min_samples_split=25, n_estimators=165; total time=   0.5s\n",
            "[CV] END max_depth=5, max_features=0.5000637862320251, min_samples_leaf=7, min_samples_split=25, n_estimators=165; total time=   0.4s\n",
            "[CV] END max_depth=5, max_features=0.5000637862320251, min_samples_leaf=7, min_samples_split=25, n_estimators=165; total time=   0.4s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=1, max_features=0.999, min_samples_leaf=10, min_samples_split=25, n_estimators=95; total time=   0.2s\n",
            "[CV] END max_depth=1, max_features=0.999, min_samples_leaf=10, min_samples_split=25, n_estimators=95; total time=   0.2s\n",
            "[CV] END max_depth=1, max_features=0.999, min_samples_leaf=10, min_samples_split=25, n_estimators=95; total time=   0.2s\n",
            "[CV] END max_depth=1, max_features=0.999, min_samples_leaf=10, min_samples_split=25, n_estimators=95; total time=   0.2s\n",
            "[CV] END max_depth=1, max_features=0.999, min_samples_leaf=10, min_samples_split=25, n_estimators=95; total time=   0.2s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=4, min_samples_split=21, n_estimators=183; total time=   0.4s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=4, min_samples_split=21, n_estimators=183; total time=   0.4s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=4, min_samples_split=21, n_estimators=183; total time=   0.5s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=4, min_samples_split=21, n_estimators=183; total time=   0.5s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=4, min_samples_split=21, n_estimators=183; total time=   0.5s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=10, max_features=0.1, min_samples_leaf=4, min_samples_split=25, n_estimators=189; total time=   0.3s\n",
            "[CV] END max_depth=10, max_features=0.1, min_samples_leaf=4, min_samples_split=25, n_estimators=189; total time=   0.3s\n",
            "[CV] END max_depth=10, max_features=0.1, min_samples_leaf=4, min_samples_split=25, n_estimators=189; total time=   0.3s\n",
            "[CV] END max_depth=10, max_features=0.1, min_samples_leaf=4, min_samples_split=25, n_estimators=189; total time=   0.3s\n",
            "[CV] END max_depth=10, max_features=0.1, min_samples_leaf=4, min_samples_split=25, n_estimators=189; total time=   0.3s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=4, max_features=0.1, min_samples_leaf=7, min_samples_split=8, n_estimators=153; total time=   0.3s\n",
            "[CV] END max_depth=4, max_features=0.1, min_samples_leaf=7, min_samples_split=8, n_estimators=153; total time=   0.3s\n",
            "[CV] END max_depth=4, max_features=0.1, min_samples_leaf=7, min_samples_split=8, n_estimators=153; total time=   0.3s\n",
            "[CV] END max_depth=4, max_features=0.1, min_samples_leaf=7, min_samples_split=8, n_estimators=153; total time=   0.3s\n",
            "[CV] END max_depth=4, max_features=0.1, min_samples_leaf=7, min_samples_split=8, n_estimators=153; total time=   0.3s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=8, max_features=0.1, min_samples_leaf=7, min_samples_split=2, n_estimators=155; total time=   0.2s\n",
            "[CV] END max_depth=8, max_features=0.1, min_samples_leaf=7, min_samples_split=2, n_estimators=155; total time=   0.2s\n",
            "[CV] END max_depth=8, max_features=0.1, min_samples_leaf=7, min_samples_split=2, n_estimators=155; total time=   0.2s\n",
            "[CV] END max_depth=8, max_features=0.1, min_samples_leaf=7, min_samples_split=2, n_estimators=155; total time=   0.2s\n",
            "[CV] END max_depth=8, max_features=0.1, min_samples_leaf=7, min_samples_split=2, n_estimators=155; total time=   0.2s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=8, min_samples_split=25, n_estimators=146; total time=   0.3s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=8, min_samples_split=25, n_estimators=146; total time=   0.3s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=8, min_samples_split=25, n_estimators=146; total time=   0.3s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=8, min_samples_split=25, n_estimators=146; total time=   0.3s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=8, min_samples_split=25, n_estimators=146; total time=   0.2s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=8, min_samples_split=18, n_estimators=141; total time=   0.2s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=8, min_samples_split=18, n_estimators=141; total time=   0.2s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=8, min_samples_split=18, n_estimators=141; total time=   0.2s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=8, min_samples_split=18, n_estimators=141; total time=   0.2s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=8, min_samples_split=18, n_estimators=141; total time=   0.2s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=8, min_samples_split=25, n_estimators=154; total time=   0.2s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=8, min_samples_split=25, n_estimators=154; total time=   0.2s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=8, min_samples_split=25, n_estimators=154; total time=   0.2s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=8, min_samples_split=25, n_estimators=154; total time=   0.2s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=8, min_samples_split=25, n_estimators=154; total time=   0.2s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=13, max_features=0.1, min_samples_leaf=9, min_samples_split=25, n_estimators=150; total time=   0.3s\n",
            "[CV] END max_depth=13, max_features=0.1, min_samples_leaf=9, min_samples_split=25, n_estimators=150; total time=   0.3s\n",
            "[CV] END max_depth=13, max_features=0.1, min_samples_leaf=9, min_samples_split=25, n_estimators=150; total time=   0.3s\n",
            "[CV] END max_depth=13, max_features=0.1, min_samples_leaf=9, min_samples_split=25, n_estimators=150; total time=   0.3s\n",
            "[CV] END max_depth=13, max_features=0.1, min_samples_leaf=9, min_samples_split=25, n_estimators=150; total time=   0.3s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=20, max_features=0.999, min_samples_leaf=7, min_samples_split=25, n_estimators=125; total time=   0.2s\n",
            "[CV] END max_depth=20, max_features=0.999, min_samples_leaf=7, min_samples_split=25, n_estimators=125; total time=   0.2s\n",
            "[CV] END max_depth=20, max_features=0.999, min_samples_leaf=7, min_samples_split=25, n_estimators=125; total time=   0.2s\n",
            "[CV] END max_depth=20, max_features=0.999, min_samples_leaf=7, min_samples_split=25, n_estimators=125; total time=   0.2s\n",
            "[CV] END max_depth=20, max_features=0.999, min_samples_leaf=7, min_samples_split=25, n_estimators=125; total time=   0.2s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=10, min_samples_split=2, n_estimators=250; total time=   0.6s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=10, min_samples_split=2, n_estimators=250; total time=   0.6s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=10, min_samples_split=2, n_estimators=250; total time=   0.6s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=10, min_samples_split=2, n_estimators=250; total time=   0.6s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=10, min_samples_split=2, n_estimators=250; total time=   0.6s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=8, min_samples_split=2, n_estimators=190; total time=   0.2s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=8, min_samples_split=2, n_estimators=190; total time=   0.2s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=8, min_samples_split=2, n_estimators=190; total time=   0.2s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=8, min_samples_split=2, n_estimators=190; total time=   0.2s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=8, min_samples_split=2, n_estimators=190; total time=   0.2s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=8, min_samples_split=25, n_estimators=164; total time=   0.3s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=8, min_samples_split=25, n_estimators=164; total time=   0.2s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=8, min_samples_split=25, n_estimators=164; total time=   0.3s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=8, min_samples_split=25, n_estimators=164; total time=   0.3s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=8, min_samples_split=25, n_estimators=164; total time=   0.2s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=8, min_samples_split=25, n_estimators=168; total time=   0.2s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=8, min_samples_split=25, n_estimators=168; total time=   0.2s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=8, min_samples_split=25, n_estimators=168; total time=   0.2s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=8, min_samples_split=25, n_estimators=168; total time=   0.2s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=8, min_samples_split=25, n_estimators=168; total time=   0.2s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=3, min_samples_split=25, n_estimators=131; total time=   0.2s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=3, min_samples_split=25, n_estimators=131; total time=   0.2s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=3, min_samples_split=25, n_estimators=131; total time=   0.2s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=3, min_samples_split=25, n_estimators=131; total time=   0.2s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=3, min_samples_split=25, n_estimators=131; total time=   0.2s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=6, min_samples_split=25, n_estimators=250; total time=   0.4s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=6, min_samples_split=25, n_estimators=250; total time=   0.4s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=6, min_samples_split=25, n_estimators=250; total time=   0.4s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=6, min_samples_split=25, n_estimators=250; total time=   0.4s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=6, min_samples_split=25, n_estimators=250; total time=   0.4s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=3, min_samples_split=25, n_estimators=140; total time=   0.2s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=3, min_samples_split=25, n_estimators=140; total time=   0.2s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=3, min_samples_split=25, n_estimators=140; total time=   0.3s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=3, min_samples_split=25, n_estimators=140; total time=   0.3s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=3, min_samples_split=25, n_estimators=140; total time=   0.3s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=1, max_features=0.29037116208634506, min_samples_leaf=8, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=1, max_features=0.29037116208634506, min_samples_leaf=8, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=1, max_features=0.29037116208634506, min_samples_leaf=8, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=1, max_features=0.29037116208634506, min_samples_leaf=8, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=1, max_features=0.29037116208634506, min_samples_leaf=8, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=7, min_samples_split=2, n_estimators=208; total time=   0.4s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=7, min_samples_split=2, n_estimators=208; total time=   0.4s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=7, min_samples_split=2, n_estimators=208; total time=   0.4s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=7, min_samples_split=2, n_estimators=208; total time=   0.4s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=7, min_samples_split=2, n_estimators=208; total time=   0.4s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=9, max_features=0.1, min_samples_leaf=6, min_samples_split=2, n_estimators=250; total time=   0.4s\n",
            "[CV] END max_depth=9, max_features=0.1, min_samples_leaf=6, min_samples_split=2, n_estimators=250; total time=   0.4s\n",
            "[CV] END max_depth=9, max_features=0.1, min_samples_leaf=6, min_samples_split=2, n_estimators=250; total time=   0.4s\n",
            "[CV] END max_depth=9, max_features=0.1, min_samples_leaf=6, min_samples_split=2, n_estimators=250; total time=   0.4s\n",
            "[CV] END max_depth=9, max_features=0.1, min_samples_leaf=6, min_samples_split=2, n_estimators=250; total time=   0.4s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=3, max_features=0.1, min_samples_leaf=5, min_samples_split=25, n_estimators=250; total time=   0.5s\n",
            "[CV] END max_depth=3, max_features=0.1, min_samples_leaf=5, min_samples_split=25, n_estimators=250; total time=   0.5s\n",
            "[CV] END max_depth=3, max_features=0.1, min_samples_leaf=5, min_samples_split=25, n_estimators=250; total time=   0.5s\n",
            "[CV] END max_depth=3, max_features=0.1, min_samples_leaf=5, min_samples_split=25, n_estimators=250; total time=   0.5s\n",
            "[CV] END max_depth=3, max_features=0.1, min_samples_leaf=5, min_samples_split=25, n_estimators=250; total time=   0.4s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=1, max_features=0.11542019014088445, min_samples_leaf=7, min_samples_split=25, n_estimators=205; total time=   0.4s\n",
            "[CV] END max_depth=1, max_features=0.11542019014088445, min_samples_leaf=7, min_samples_split=25, n_estimators=205; total time=   0.4s\n",
            "[CV] END max_depth=1, max_features=0.11542019014088445, min_samples_leaf=7, min_samples_split=25, n_estimators=205; total time=   0.4s\n",
            "[CV] END max_depth=1, max_features=0.11542019014088445, min_samples_leaf=7, min_samples_split=25, n_estimators=205; total time=   0.4s\n",
            "[CV] END max_depth=1, max_features=0.11542019014088445, min_samples_leaf=7, min_samples_split=25, n_estimators=205; total time=   0.4s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=8, min_samples_split=25, n_estimators=50; total time=   0.0s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=8, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=8, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=8, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=8, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=7, min_samples_split=25, n_estimators=220; total time=   0.3s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=7, min_samples_split=25, n_estimators=220; total time=   0.4s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=7, min_samples_split=25, n_estimators=220; total time=   0.4s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=7, min_samples_split=25, n_estimators=220; total time=   0.4s\n",
            "[CV] END max_depth=20, max_features=0.1, min_samples_leaf=7, min_samples_split=25, n_estimators=220; total time=   0.4s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=1, max_features=0.6908950558670599, min_samples_leaf=5, min_samples_split=2, n_estimators=153; total time=   0.2s\n",
            "[CV] END max_depth=1, max_features=0.6908950558670599, min_samples_leaf=5, min_samples_split=2, n_estimators=153; total time=   0.2s\n",
            "[CV] END max_depth=1, max_features=0.6908950558670599, min_samples_leaf=5, min_samples_split=2, n_estimators=153; total time=   0.2s\n",
            "[CV] END max_depth=1, max_features=0.6908950558670599, min_samples_leaf=5, min_samples_split=2, n_estimators=153; total time=   0.2s\n",
            "[CV] END max_depth=1, max_features=0.6908950558670599, min_samples_leaf=5, min_samples_split=2, n_estimators=153; total time=   0.2s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=3, min_samples_split=25, n_estimators=139; total time=   0.3s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=3, min_samples_split=25, n_estimators=139; total time=   0.3s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=3, min_samples_split=25, n_estimators=139; total time=   0.3s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=3, min_samples_split=25, n_estimators=139; total time=   0.3s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=3, min_samples_split=25, n_estimators=139; total time=   0.3s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=1, max_features=0.999, min_samples_leaf=8, min_samples_split=25, n_estimators=250; total time=   0.4s\n",
            "[CV] END max_depth=1, max_features=0.999, min_samples_leaf=8, min_samples_split=25, n_estimators=250; total time=   0.4s\n",
            "[CV] END max_depth=1, max_features=0.999, min_samples_leaf=8, min_samples_split=25, n_estimators=250; total time=   0.4s\n",
            "[CV] END max_depth=1, max_features=0.999, min_samples_leaf=8, min_samples_split=25, n_estimators=250; total time=   0.4s\n",
            "[CV] END max_depth=1, max_features=0.999, min_samples_leaf=8, min_samples_split=25, n_estimators=250; total time=   0.4s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=3, max_features=0.2790793762995375, min_samples_leaf=9, min_samples_split=25, n_estimators=152; total time=   0.2s\n",
            "[CV] END max_depth=3, max_features=0.2790793762995375, min_samples_leaf=9, min_samples_split=25, n_estimators=152; total time=   0.2s\n",
            "[CV] END max_depth=3, max_features=0.2790793762995375, min_samples_leaf=9, min_samples_split=25, n_estimators=152; total time=   0.2s\n",
            "[CV] END max_depth=3, max_features=0.2790793762995375, min_samples_leaf=9, min_samples_split=25, n_estimators=152; total time=   0.2s\n",
            "[CV] END max_depth=3, max_features=0.2790793762995375, min_samples_leaf=9, min_samples_split=25, n_estimators=152; total time=   0.2s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=1, max_features=0.999, min_samples_leaf=4, min_samples_split=25, n_estimators=250; total time=   0.4s\n",
            "[CV] END max_depth=1, max_features=0.999, min_samples_leaf=4, min_samples_split=25, n_estimators=250; total time=   0.4s\n",
            "[CV] END max_depth=1, max_features=0.999, min_samples_leaf=4, min_samples_split=25, n_estimators=250; total time=   0.4s\n",
            "[CV] END max_depth=1, max_features=0.999, min_samples_leaf=4, min_samples_split=25, n_estimators=250; total time=   0.3s\n",
            "[CV] END max_depth=1, max_features=0.999, min_samples_leaf=4, min_samples_split=25, n_estimators=250; total time=   0.4s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=6, min_samples_split=2, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=6, min_samples_split=2, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=6, min_samples_split=2, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=6, min_samples_split=2, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=1, max_features=0.1, min_samples_leaf=6, min_samples_split=2, n_estimators=50; total time=   0.1s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=14, max_features=0.999, min_samples_leaf=10, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=14, max_features=0.999, min_samples_leaf=10, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=14, max_features=0.999, min_samples_leaf=10, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=14, max_features=0.999, min_samples_leaf=10, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=14, max_features=0.999, min_samples_leaf=10, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=1, max_features=0.1121069814295376, min_samples_leaf=10, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=1, max_features=0.1121069814295376, min_samples_leaf=10, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=1, max_features=0.1121069814295376, min_samples_leaf=10, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=1, max_features=0.1121069814295376, min_samples_leaf=10, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=1, max_features=0.1121069814295376, min_samples_leaf=10, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=20, max_features=0.10044957030986605, min_samples_leaf=9, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=20, max_features=0.10044957030986605, min_samples_leaf=9, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=20, max_features=0.10044957030986605, min_samples_leaf=9, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=20, max_features=0.10044957030986605, min_samples_leaf=9, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=20, max_features=0.10044957030986605, min_samples_leaf=9, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=1, max_features=0.3174607510723978, min_samples_leaf=9, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=1, max_features=0.3174607510723978, min_samples_leaf=9, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=1, max_features=0.3174607510723978, min_samples_leaf=9, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=1, max_features=0.3174607510723978, min_samples_leaf=9, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=1, max_features=0.3174607510723978, min_samples_leaf=9, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=3, max_features=0.1, min_samples_leaf=10, min_samples_split=7, n_estimators=50; total time=   0.0s\n",
            "[CV] END max_depth=3, max_features=0.1, min_samples_leaf=10, min_samples_split=7, n_estimators=50; total time=   0.0s\n",
            "[CV] END max_depth=3, max_features=0.1, min_samples_leaf=10, min_samples_split=7, n_estimators=50; total time=   0.0s\n",
            "[CV] END max_depth=3, max_features=0.1, min_samples_leaf=10, min_samples_split=7, n_estimators=50; total time=   0.0s\n",
            "[CV] END max_depth=3, max_features=0.1, min_samples_leaf=10, min_samples_split=7, n_estimators=50; total time=   0.0s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=20, max_features=0.1334560372989192, min_samples_leaf=5, min_samples_split=21, n_estimators=242; total time=   0.3s\n",
            "[CV] END max_depth=20, max_features=0.1334560372989192, min_samples_leaf=5, min_samples_split=21, n_estimators=242; total time=   0.3s\n",
            "[CV] END max_depth=20, max_features=0.1334560372989192, min_samples_leaf=5, min_samples_split=21, n_estimators=242; total time=   0.3s\n",
            "[CV] END max_depth=20, max_features=0.1334560372989192, min_samples_leaf=5, min_samples_split=21, n_estimators=242; total time=   0.3s\n",
            "[CV] END max_depth=20, max_features=0.1334560372989192, min_samples_leaf=5, min_samples_split=21, n_estimators=242; total time=   0.4s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=13, max_features=0.1, min_samples_leaf=1, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=13, max_features=0.1, min_samples_leaf=1, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=13, max_features=0.1, min_samples_leaf=1, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=13, max_features=0.1, min_samples_leaf=1, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=13, max_features=0.1, min_samples_leaf=1, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
            "[CV] END max_depth=10, max_features=0.1514209249726602, min_samples_leaf=8, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=10, max_features=0.1514209249726602, min_samples_leaf=8, min_samples_split=25, n_estimators=50; total time=   0.1s\n",
            "[CV] END max_depth=10, max_features=0.1514209249726602, min_samples_leaf=8, min_samples_split=25, n_estimators=50; total time=   0.0s\n",
            "[CV] END max_depth=10, max_features=0.1514209249726602, min_samples_leaf=8, min_samples_split=25, n_estimators=50; total time=   0.0s\n",
            "[CV] END max_depth=10, max_features=0.1514209249726602, min_samples_leaf=8, min_samples_split=25, n_estimators=50; total time=   0.0s\n",
            "Best Hyperparameters: OrderedDict([('max_depth', 10), ('max_features', 0.1514209249726602), ('min_samples_leaf', 8), ('min_samples_split', 25), ('n_estimators', 50)])\n"
          ]
        }
      ],
      "source": [
        "# Define the hyperparameters and their search spaces\n",
        "param_dist = {\n",
        "    'n_estimators': (50, 250),\n",
        "    'max_depth': (1, 20),\n",
        "    'min_samples_split': (2, 25),\n",
        "    'min_samples_leaf': (1, 10),\n",
        "    'max_features': (0.1, 0.999),\n",
        "}\n",
        "\n",
        "# Create BayesSearchCV\n",
        "search = BayesSearchCV(estimator=rf_regressor, search_spaces=param_dist, scoring='neg_mean_squared_error', n_iter=50, cv=5, n_jobs=-1, verbose=2)\n",
        "\n",
        "# Fit the model\n",
        "search.fit(X_train, y_train)\n",
        "\n",
        "# Get the best hyperparameters\n",
        "best_params = search.best_params_\n",
        "print(\"Best Hyperparameters:\", best_params)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "#### 2. Grid Search"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 652,
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Fitting 5 folds for each of 432 candidates, totalling 2160 fits\n",
            "Best Hyperparameters: {'bootstrap': True, 'max_depth': None, 'max_features': 'sqrt', 'min_samples_leaf': 4, 'min_samples_split': 10, 'n_estimators': 100}\n"
          ]
        }
      ],
      "source": [
        "# First tour\n",
        "\n",
        "param_grid = {\n",
        "    \"n_estimators\": [50, 100, 150],          \n",
        "    \"max_features\": [\"sqrt\", \"log2\"], \n",
        "    \"max_depth\": [None, 10, 20, 30],           \n",
        "    \"min_samples_split\": [2, 5, 10],           \n",
        "    \"min_samples_leaf\": [1, 2, 4],             \n",
        "    \"bootstrap\": [True, False]\n",
        "}\n",
        "\n",
        "# Create the Random Forest Regressor\n",
        "rf_regressor = RandomForestRegressor(random_state=42)\n",
        "\n",
        "# Create GridSearchCV with verbose\n",
        "search = GridSearchCV(estimator=rf_regressor, param_grid=param_grid, scoring='neg_mean_squared_error', cv=5, n_jobs=-1, verbose=1)\n",
        "\n",
        "# Fit the model\n",
        "search.fit(X_train, y_train)\n",
        "\n",
        "# Get the best hyperparameters\n",
        "best_params = search.best_params_\n",
        "print(\"Best Hyperparameters:\", best_params)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "### Evaluation"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 653,
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Mean Squared Error on Test Set: 93.582\n",
            "R2 Score on Test Set: 16.642%\n"
          ]
        }
      ],
      "source": [
        "# Get the best model\n",
        "best_rf_model = search.best_estimator_\n",
        "\n",
        "# Make predictions on the test set\n",
        "y_pred = best_rf_model.predict(X_test)\n",
        "\n",
        "# Evaluate the model\n",
        "mse = mean_squared_error(y_test, y_pred)\n",
        "print(f'Mean Squared Error on Test Set: {round(mse, 3)}')\n",
        "\n",
        "print(f\"R2 Score on Test Set: {round(100 * r2_score(y_test, y_pred), 3)}%\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "## 4.3. XGBoost"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 654,
      "metadata": {},
      "outputs": [
        {
          "data": {
            "text/html": [
              "<style>#sk-container-id-46 {color: black;background-color: white;}#sk-container-id-46 pre{padding: 0;}#sk-container-id-46 div.sk-toggleable {background-color: white;}#sk-container-id-46 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-46 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-46 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-46 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-46 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-46 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-46 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-46 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-46 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-46 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-46 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-46 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-46 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-46 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-46 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-46 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-46 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-46 div.sk-item {position: relative;z-index: 1;}#sk-container-id-46 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-46 div.sk-item::before, #sk-container-id-46 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-46 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-46 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-46 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-46 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-46 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-46 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-46 div.sk-label-container {text-align: center;}#sk-container-id-46 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-46 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-46\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>XGBRegressor(base_score=0.5, booster=&#x27;gbtree&#x27;, callbacks=None,\n",
              "             colsample_bylevel=1, colsample_bynode=1, colsample_bytree=1,\n",
              "             early_stopping_rounds=None, enable_categorical=False,\n",
              "             eval_metric=None, feature_types=None, gamma=0, gpu_id=-1,\n",
              "             grow_policy=&#x27;depthwise&#x27;, importance_type=None,\n",
              "             interaction_constraints=&#x27;&#x27;, learning_rate=0.300000012, max_bin=256,\n",
              "             max_cat_threshold=64, max_cat_to_onehot=4, max_delta_step=0,\n",
              "             max_depth=6, max_leaves=0, min_child_weight=1, missing=nan,\n",
              "             monotone_constraints=&#x27;()&#x27;, n_estimators=100, n_jobs=0,\n",
              "             num_parallel_tree=1, predictor=&#x27;auto&#x27;, random_state=42, ...)</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-46\" type=\"checkbox\" checked><label for=\"sk-estimator-id-46\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">XGBRegressor</label><div class=\"sk-toggleable__content\"><pre>XGBRegressor(base_score=0.5, booster=&#x27;gbtree&#x27;, callbacks=None,\n",
              "             colsample_bylevel=1, colsample_bynode=1, colsample_bytree=1,\n",
              "             early_stopping_rounds=None, enable_categorical=False,\n",
              "             eval_metric=None, feature_types=None, gamma=0, gpu_id=-1,\n",
              "             grow_policy=&#x27;depthwise&#x27;, importance_type=None,\n",
              "             interaction_constraints=&#x27;&#x27;, learning_rate=0.300000012, max_bin=256,\n",
              "             max_cat_threshold=64, max_cat_to_onehot=4, max_delta_step=0,\n",
              "             max_depth=6, max_leaves=0, min_child_weight=1, missing=nan,\n",
              "             monotone_constraints=&#x27;()&#x27;, n_estimators=100, n_jobs=0,\n",
              "             num_parallel_tree=1, predictor=&#x27;auto&#x27;, random_state=42, ...)</pre></div></div></div></div></div>"
            ],
            "text/plain": [
              "XGBRegressor(base_score=0.5, booster='gbtree', callbacks=None,\n",
              "             colsample_bylevel=1, colsample_bynode=1, colsample_bytree=1,\n",
              "             early_stopping_rounds=None, enable_categorical=False,\n",
              "             eval_metric=None, feature_types=None, gamma=0, gpu_id=-1,\n",
              "             grow_policy='depthwise', importance_type=None,\n",
              "             interaction_constraints='', learning_rate=0.300000012, max_bin=256,\n",
              "             max_cat_threshold=64, max_cat_to_onehot=4, max_delta_step=0,\n",
              "             max_depth=6, max_leaves=0, min_child_weight=1, missing=nan,\n",
              "             monotone_constraints='()', n_estimators=100, n_jobs=0,\n",
              "             num_parallel_tree=1, predictor='auto', random_state=42, ...)"
            ]
          },
          "execution_count": 654,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "xgb_regressor = XGBRegressor(random_state=42)\n",
        "xgb_regressor.fit(X_train, y_train)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 655,
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "MSE Train: 4.477583529150411e-07\n",
            "MSE TEST: 81.20227402488236\n",
            "R2 Train: 0.9999999972600819\n",
            "R2 TEST: 0.27669496243833946\n"
          ]
        }
      ],
      "source": [
        "# Prediction\n",
        "y_train_pred = xgb_regressor.predict(X_train)\n",
        "y_test_pred = xgb_regressor.predict(X_test)\n",
        "\n",
        "# Calculation of Mean Squared Error (MSE)\n",
        "print(\"MSE Train:\", mean_squared_error(y_train,y_train_pred))\n",
        "print(\"MSE TEST:\", mean_squared_error(y_test,y_test_pred))\n",
        "\n",
        "print(\"R2 Train:\", r2_score(y_train,y_train_pred))\n",
        "print(\"R2 TEST:\", r2_score(y_test,y_test_pred))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 656,
      "metadata": {},
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "plot_importance(xgb_regressor)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "### Feature Selection"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "* Based on feature importance\n",
        "* All subset of features (sorted by importance) are evaluated to get the best model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 657,
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Threshold: 0.0005, #Features: 32, MSE: 76.51\n",
            "Threshold: 0.0005, #Features: 31, MSE: 76.51\n",
            "Threshold: 0.0008, #Features: 30, MSE: 75.00\n",
            "Threshold: 0.0013, #Features: 29, MSE: 76.55\n",
            "Threshold: 0.0014, #Features: 28, MSE: 76.53\n",
            "Threshold: 0.0031, #Features: 27, MSE: 75.66\n",
            "Threshold: 0.0033, #Features: 26, MSE: 76.29\n",
            "Threshold: 0.0036, #Features: 25, MSE: 70.29\n",
            "Threshold: 0.0052, #Features: 24, MSE: 70.29\n",
            "Threshold: 0.0063, #Features: 23, MSE: 71.08\n",
            "Threshold: 0.0092, #Features: 22, MSE: 71.03\n",
            "Threshold: 0.0095, #Features: 21, MSE: 67.95\n",
            "Threshold: 0.0120, #Features: 20, MSE: 78.77\n",
            "Threshold: 0.0122, #Features: 19, MSE: 77.99\n",
            "Threshold: 0.0127, #Features: 18, MSE: 79.43\n",
            "Threshold: 0.0142, #Features: 17, MSE: 53.43\n",
            "Threshold: 0.0167, #Features: 16, MSE: 55.52\n",
            "Threshold: 0.0169, #Features: 15, MSE: 55.42\n",
            "Threshold: 0.0176, #Features: 14, MSE: 59.99\n",
            "Threshold: 0.0212, #Features: 13, MSE: 51.05\n",
            "Threshold: 0.0229, #Features: 12, MSE: 53.11\n",
            "Threshold: 0.0233, #Features: 11, MSE: 40.98\n",
            "Threshold: 0.0263, #Features: 10, MSE: 45.22\n",
            "Threshold: 0.0293, #Features: 9, MSE: 47.64\n",
            "Threshold: 0.0300, #Features: 8, MSE: 60.68\n",
            "Threshold: 0.0416, #Features: 7, MSE: 94.21\n",
            "Threshold: 0.0499, #Features: 6, MSE: 125.44\n",
            "Threshold: 0.0734, #Features: 5, MSE: 130.31\n",
            "Threshold: 0.0746, #Features: 4, MSE: 132.54\n",
            "Threshold: 0.0870, #Features: 3, MSE: 204.48\n",
            "Threshold: 0.0917, #Features: 2, MSE: 128.31\n",
            "Threshold: 0.2821, #Features: 1, MSE: 112.37\n",
            "\n",
            "Best threshold: 0.023293254896998405\n",
            "Best MSE: 40.984\n"
          ]
        }
      ],
      "source": [
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)\n",
        "\n",
        "xgb_regressor = XGBRegressor(random_state=42)\n",
        "xgb_regressor.fit(X_train, y_train)\n",
        "\n",
        "# Select most important features & re-train\n",
        "thresholds = np.sort(xgb_regressor.feature_importances_)\n",
        "best_threshold = 0\n",
        "best_mse = 1000\n",
        "best_xgb_model = xgb_regressor\n",
        "\n",
        "# Set each feature as a threshold and evaluate resulting model on the test set\n",
        "for thresh in thresholds:\n",
        "    # Select features using threshold\n",
        "    selection = SelectFromModel(xgb_regressor, threshold=thresh, prefit=True)\n",
        "    select_X_train = selection.transform(X_train)\n",
        "    select_X_test = selection.transform(X_test)\n",
        "\n",
        "    # Train model with selected features\n",
        "    eval_set = [(select_X_test, y_test)]\n",
        "    selection_model = XGBRegressor(early_stopping_rounds=20, eval_metric=\"rmse\")\n",
        "    selection_model.fit(select_X_train, y_train, eval_set=eval_set, verbose=0)\n",
        "    \n",
        "    # Evaluate the trained model\n",
        "    y_test_pred = selection_model.predict(select_X_test)\n",
        "    mse = mean_squared_error(y_test,y_test_pred)\n",
        "    print(\"Threshold: %.4f, #Features: %d, MSE: %.2f\" % (thresh, select_X_train.shape[1], mse))\n",
        "\n",
        "    if (mse < best_mse):\n",
        "        best_mse = mse\n",
        "        best_threshold = thresh\n",
        "        best_xgb_model = selection_model\n",
        "\n",
        "print()\n",
        "print(f\"Best threshold: {best_threshold}\")\n",
        "print(f\"Best MSE: {round(best_mse, 3)}\")\n",
        "\n",
        "# Select features based on the best threshold\n",
        "selection = SelectFromModel(xgb_regressor, threshold=best_threshold, prefit=True)\n",
        "select_X_train = selection.transform(X_train)\n",
        "select_X_test = selection.transform(X_test)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "### Evaluation"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 658,
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Mean Squared Error on Test Set: 40.98355483588821\n",
            "R2 Score on Test Set: 63.494%\n"
          ]
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "# Make predictions on the test set\n",
        "y_pred = best_xgb_model.predict(select_X_test)\n",
        "\n",
        "# Evaluate the model\n",
        "mse = mean_squared_error(y_test, y_pred)\n",
        "print(f'Mean Squared Error on Test Set: {mse}')\n",
        "print(f\"R2 Score on Test Set: {round(100 * r2_score(y_test, y_pred), 3)}%\")\n",
        "\n",
        "plot_importance(best_xgb_model)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "## 4.4 Neural Network"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 659,
      "metadata": {},
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "2024-01-19 13:10:43.202100: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 AVX512F AVX512_VNNI FMA\n",
            "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n"
          ]
        }
      ],
      "source": [
        "from keras.models import Sequential\n",
        "from keras.layers import Dense\n",
        "from keras.callbacks import EarlyStopping\n",
        "from sklearn.preprocessing import StandardScaler"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 660,
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1/100\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "2024-01-19 13:11:07.550655: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 AVX512F AVX512_VNNI FMA\n",
            "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "3/3 [==============================] - 1s 131ms/step - loss: 8456.4150 - mae: 90.8753 - val_loss: 8839.5020 - val_mae: 93.8990\n",
            "Epoch 2/100\n",
            "3/3 [==============================] - 0s 17ms/step - loss: 8291.5625 - mae: 89.9538 - val_loss: 8674.0479 - val_mae: 93.0108\n",
            "Epoch 3/100\n",
            "3/3 [==============================] - 0s 17ms/step - loss: 8107.1357 - mae: 88.8970 - val_loss: 8474.0332 - val_mae: 91.9240\n",
            "Epoch 4/100\n",
            "3/3 [==============================] - 0s 17ms/step - loss: 7888.2339 - mae: 87.6287 - val_loss: 8218.3701 - val_mae: 90.5136\n",
            "Epoch 5/100\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 7604.1113 - mae: 85.9571 - val_loss: 7883.5654 - val_mae: 88.6263\n",
            "Epoch 6/100\n",
            "3/3 [==============================] - 0s 18ms/step - loss: 7227.9409 - mae: 83.6854 - val_loss: 7449.6509 - val_mae: 86.1078\n",
            "Epoch 7/100\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 6743.6343 - mae: 80.6169 - val_loss: 6898.6768 - val_mae: 82.7789\n",
            "Epoch 8/100\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 6142.0488 - mae: 76.5875 - val_loss: 6211.1699 - val_mae: 78.3876\n",
            "Epoch 9/100\n",
            "3/3 [==============================] - 0s 17ms/step - loss: 5398.2710 - mae: 71.4818 - val_loss: 5382.5894 - val_mae: 72.6626\n",
            "Epoch 10/100\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 4538.3315 - mae: 65.0076 - val_loss: 4436.3242 - val_mae: 65.3600\n",
            "Epoch 11/100\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 3587.0906 - mae: 56.3923 - val_loss: 3435.5293 - val_mae: 56.3164\n",
            "Epoch 12/100\n",
            "3/3 [==============================] - 0s 17ms/step - loss: 2756.8989 - mae: 48.7652 - val_loss: 2500.3997 - val_mae: 45.7047\n",
            "Epoch 13/100\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 2088.5979 - mae: 40.9453 - val_loss: 1802.8777 - val_mae: 37.2773\n",
            "Epoch 14/100\n",
            "3/3 [==============================] - 0s 17ms/step - loss: 1759.6078 - mae: 36.2715 - val_loss: 1403.9043 - val_mae: 32.3238\n",
            "Epoch 15/100\n",
            "3/3 [==============================] - 0s 17ms/step - loss: 1561.0305 - mae: 33.0516 - val_loss: 1236.0414 - val_mae: 29.4063\n",
            "Epoch 16/100\n",
            "3/3 [==============================] - 0s 17ms/step - loss: 1486.0288 - mae: 30.4764 - val_loss: 1158.4561 - val_mae: 28.0933\n",
            "Epoch 17/100\n",
            "3/3 [==============================] - 0s 17ms/step - loss: 1329.1522 - mae: 28.4303 - val_loss: 1149.9026 - val_mae: 27.8518\n",
            "Epoch 18/100\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 1148.1174 - mae: 26.5926 - val_loss: 1191.2842 - val_mae: 28.4996\n",
            "Epoch 19/100\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 1001.2420 - mae: 25.1535 - val_loss: 1255.3250 - val_mae: 29.8819\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x13f943e80>"
            ]
          },
          "execution_count": 660,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "NN_model = Sequential()\n",
        "\n",
        "# Input layer\n",
        "NN_model.add(Dense(256, input_shape=(X_train.shape[1],), activation='relu'))\n",
        "\n",
        "# Hidden layers\n",
        "NN_model.add(Dense(128, activation='relu'))\n",
        "NN_model.add(Dense(64, activation='relu'))\n",
        "\n",
        "# Output layers\n",
        "NN_model.add(Dense(1))\n",
        "\n",
        "# Set up early stopping to avoid overfitting\n",
        "early_stopping = EarlyStopping(monitor='val_loss', patience=2)\n",
        "\n",
        "# Standardize the data (optional but often recommended)\n",
        "scaler = StandardScaler()\n",
        "X_train_scaled = scaler.fit_transform(X_train)\n",
        "X_test_scaled = scaler.transform(X_test)\n",
        "\n",
        "NN_model.compile(optimizer='adam', loss='mean_squared_error', metrics=['mae'])\n",
        "NN_model.fit(X_train_scaled, y_train, epochs=100, batch_size=32, validation_split = 0.2, callbacks=[early_stopping])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 661,
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "1/1 [==============================] - 0s 93ms/step - loss: 613.4781 - mae: 19.9888\n",
            "Test Mean Squared Error: 613.48\n",
            "Test Mean Absolute Error: 19.99\n"
          ]
        }
      ],
      "source": [
        "# Evaluate the model on the test set\n",
        "test_loss, test_mae = NN_model.evaluate(X_test_scaled, y_test)\n",
        "test_loss = round(test_loss, 2)\n",
        "test_mae = round(test_mae, 2)\n",
        "print(f'Test Mean Squared Error: {test_loss}')\n",
        "print(f'Test Mean Absolute Error: {test_mae}')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "## 4.5. Linear Regression"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 662,
      "metadata": {},
      "outputs": [],
      "source": [
        "from sklearn.linear_model import LinearRegression"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 663,
      "metadata": {},
      "outputs": [
        {
          "data": {
            "text/html": [
              "<style>#sk-container-id-47 {color: black;background-color: white;}#sk-container-id-47 pre{padding: 0;}#sk-container-id-47 div.sk-toggleable {background-color: white;}#sk-container-id-47 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-47 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-47 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-47 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-47 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-47 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-47 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-47 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-47 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-47 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-47 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-47 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-47 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-47 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-47 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-47 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-47 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-47 div.sk-item {position: relative;z-index: 1;}#sk-container-id-47 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-47 div.sk-item::before, #sk-container-id-47 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-47 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-47 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-47 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-47 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-47 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-47 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-47 div.sk-label-container {text-align: center;}#sk-container-id-47 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-47 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-47\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>LinearRegression()</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-47\" type=\"checkbox\" checked><label for=\"sk-estimator-id-47\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">LinearRegression</label><div class=\"sk-toggleable__content\"><pre>LinearRegression()</pre></div></div></div></div></div>"
            ],
            "text/plain": [
              "LinearRegression()"
            ]
          },
          "execution_count": 663,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "lr_model = LinearRegression()\n",
        "\n",
        "lr_model.fit(X_train, y_train)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 664,
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "MSE Train: 103.53749058762818\n",
            "MSE TEST: 97.72888035097269\n",
            "R2 Train: 0.3664345014971212\n",
            "R2 TEST: 0.12948507511675267\n"
          ]
        }
      ],
      "source": [
        "# Prediction\n",
        "y_train_pred = lr_model.predict(X_train)\n",
        "y_test_pred = lr_model.predict(X_test)\n",
        "\n",
        "# Calculation of Mean Squared Error (MSE)\n",
        "print(\"MSE Train:\", mean_squared_error(y_train,y_train_pred))\n",
        "print(\"MSE TEST:\", mean_squared_error(y_test,y_test_pred))\n",
        "\n",
        "print(\"R2 Train:\", r2_score(y_train,y_train_pred))\n",
        "print(\"R2 TEST:\", r2_score(y_test,y_test_pred))"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "collapsed_sections": [
        "vEUofgQXqSeh",
        "Bu-Nz6tBqZ2u",
        "t_nBwH7_rM7g",
        "kqz8dG7crVux",
        "HklW_rk1rh9M",
        "qTIy5TMQqvyA",
        "5U--jc1JrtIO",
        "Ma80FZjRs4LP",
        "FaryZpiytCmI"
      ],
      "provenance": []
    },
    "kernelspec": {
      "display_name": "venv",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.10.7"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}