letter-recognition.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "c9010fbc",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.metrics import accuracy_score, classification_report\n",
    "\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "from keras import callbacks"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "0b88d4ec",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('data/letter-recognition.data', header=None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "ca66a07c",
   "metadata": {},
   "outputs": [],
   "source": [
    "df.columns = ['letter', 'x-box', 'y-box', 'width', 'high', 'onpix', 'x-bar', 'y-bar', 'x2bar', 'y2bar', 'xybar', 'x2ybr', 'xy2br', 'x-ege', 'xegvy', 'y-ege', 'yegvx']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "853bb17a",
   "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>letter</th>\n",
       "      <th>x-box</th>\n",
       "      <th>y-box</th>\n",
       "      <th>width</th>\n",
       "      <th>high</th>\n",
       "      <th>onpix</th>\n",
       "      <th>x-bar</th>\n",
       "      <th>y-bar</th>\n",
       "      <th>x2bar</th>\n",
       "      <th>y2bar</th>\n",
       "      <th>xybar</th>\n",
       "      <th>x2ybr</th>\n",
       "      <th>xy2br</th>\n",
       "      <th>x-ege</th>\n",
       "      <th>xegvy</th>\n",
       "      <th>y-ege</th>\n",
       "      <th>yegvx</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>T</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>3</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>8</td>\n",
       "      <td>13</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>6</td>\n",
       "      <td>10</td>\n",
       "      <td>8</td>\n",
       "      <td>0</td>\n",
       "      <td>8</td>\n",
       "      <td>0</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>I</td>\n",
       "      <td>5</td>\n",
       "      <td>12</td>\n",
       "      <td>3</td>\n",
       "      <td>7</td>\n",
       "      <td>2</td>\n",
       "      <td>10</td>\n",
       "      <td>5</td>\n",
       "      <td>5</td>\n",
       "      <td>4</td>\n",
       "      <td>13</td>\n",
       "      <td>3</td>\n",
       "      <td>9</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>4</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>D</td>\n",
       "      <td>4</td>\n",
       "      <td>11</td>\n",
       "      <td>6</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>10</td>\n",
       "      <td>6</td>\n",
       "      <td>2</td>\n",
       "      <td>6</td>\n",
       "      <td>10</td>\n",
       "      <td>3</td>\n",
       "      <td>7</td>\n",
       "      <td>3</td>\n",
       "      <td>7</td>\n",
       "      <td>3</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>N</td>\n",
       "      <td>7</td>\n",
       "      <td>11</td>\n",
       "      <td>6</td>\n",
       "      <td>6</td>\n",
       "      <td>3</td>\n",
       "      <td>5</td>\n",
       "      <td>9</td>\n",
       "      <td>4</td>\n",
       "      <td>6</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>10</td>\n",
       "      <td>6</td>\n",
       "      <td>10</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>G</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>6</td>\n",
       "      <td>6</td>\n",
       "      <td>6</td>\n",
       "      <td>5</td>\n",
       "      <td>9</td>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>5</td>\n",
       "      <td>10</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",
       "      <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>19995</th>\n",
       "      <td>D</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>6</td>\n",
       "      <td>6</td>\n",
       "      <td>6</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>3</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19996</th>\n",
       "      <td>C</td>\n",
       "      <td>7</td>\n",
       "      <td>10</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>9</td>\n",
       "      <td>12</td>\n",
       "      <td>9</td>\n",
       "      <td>13</td>\n",
       "      <td>2</td>\n",
       "      <td>9</td>\n",
       "      <td>3</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19997</th>\n",
       "      <td>T</td>\n",
       "      <td>6</td>\n",
       "      <td>9</td>\n",
       "      <td>6</td>\n",
       "      <td>7</td>\n",
       "      <td>5</td>\n",
       "      <td>6</td>\n",
       "      <td>11</td>\n",
       "      <td>3</td>\n",
       "      <td>7</td>\n",
       "      <td>11</td>\n",
       "      <td>9</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>12</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19998</th>\n",
       "      <td>S</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>8</td>\n",
       "      <td>7</td>\n",
       "      <td>2</td>\n",
       "      <td>6</td>\n",
       "      <td>10</td>\n",
       "      <td>6</td>\n",
       "      <td>8</td>\n",
       "      <td>1</td>\n",
       "      <td>9</td>\n",
       "      <td>5</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19999</th>\n",
       "      <td>A</td>\n",
       "      <td>4</td>\n",
       "      <td>9</td>\n",
       "      <td>6</td>\n",
       "      <td>6</td>\n",
       "      <td>2</td>\n",
       "      <td>9</td>\n",
       "      <td>5</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>8</td>\n",
       "      <td>1</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>7</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>20000 rows × 17 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      letter  x-box  y-box  width  high  onpix  x-bar  y-bar  x2bar  y2bar  \\\n",
       "0          T      2      8      3     5      1      8     13      0      6   \n",
       "1          I      5     12      3     7      2     10      5      5      4   \n",
       "2          D      4     11      6     8      6     10      6      2      6   \n",
       "3          N      7     11      6     6      3      5      9      4      6   \n",
       "4          G      2      1      3     1      1      8      6      6      6   \n",
       "...      ...    ...    ...    ...   ...    ...    ...    ...    ...    ...   \n",
       "19995      D      2      2      3     3      2      7      7      7      6   \n",
       "19996      C      7     10      8     8      4      4      8      6      9   \n",
       "19997      T      6      9      6     7      5      6     11      3      7   \n",
       "19998      S      2      3      4     2      1      8      7      2      6   \n",
       "19999      A      4      9      6     6      2      9      5      3      1   \n",
       "\n",
       "       xybar  x2ybr  xy2br  x-ege  xegvy  y-ege  yegvx  \n",
       "0          6     10      8      0      8      0      8  \n",
       "1         13      3      9      2      8      4     10  \n",
       "2         10      3      7      3      7      3      9  \n",
       "3          4      4     10      6     10      2      8  \n",
       "4          6      5      9      1      7      5     10  \n",
       "...      ...    ...    ...    ...    ...    ...    ...  \n",
       "19995      6      6      4      2      8      3      7  \n",
       "19996     12      9     13      2      9      3      7  \n",
       "19997     11      9      5      2     12      2      4  \n",
       "19998     10      6      8      1      9      5      8  \n",
       "19999      8      1      8      2      7      2      8  \n",
       "\n",
       "[20000 rows x 17 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "57f1c039",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "U    4.1%\n",
       "D    4.0%\n",
       "P    4.0%\n",
       "T    4.0%\n",
       "M    4.0%\n",
       "A    3.9%\n",
       "X    3.9%\n",
       "Y    3.9%\n",
       "N    3.9%\n",
       "Q    3.9%\n",
       "F    3.9%\n",
       "G    3.9%\n",
       "E    3.8%\n",
       "B    3.8%\n",
       "V    3.8%\n",
       "L    3.8%\n",
       "R    3.8%\n",
       "I    3.8%\n",
       "O    3.8%\n",
       "W    3.8%\n",
       "S    3.7%\n",
       "J    3.7%\n",
       "K    3.7%\n",
       "C    3.7%\n",
       "H    3.7%\n",
       "Z    3.7%\n",
       "Name: letter, dtype: object"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['letter'].value_counts(normalize=True).mul(100).round(1).astype(str) + '%'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "b762fb11",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_without_letter_column = df.loc[:, df.columns != 'letter']\n",
    "letter_column = df['letter']\n",
    "x = df_without_letter_column.values \n",
    "min_max_scaler = MinMaxScaler()\n",
    "x_scaled = min_max_scaler.fit_transform(x)\n",
    "df = pd.DataFrame(x_scaled)\n",
    "df.insert(0, 'letter', letter_column)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "b0feeb09",
   "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>letter</th>\n",
       "      <th>x-box</th>\n",
       "      <th>y-box</th>\n",
       "      <th>width</th>\n",
       "      <th>high</th>\n",
       "      <th>onpix</th>\n",
       "      <th>x-bar</th>\n",
       "      <th>y-bar</th>\n",
       "      <th>x2bar</th>\n",
       "      <th>y2bar</th>\n",
       "      <th>xybar</th>\n",
       "      <th>x2ybr</th>\n",
       "      <th>xy2br</th>\n",
       "      <th>x-ege</th>\n",
       "      <th>xegvy</th>\n",
       "      <th>y-ege</th>\n",
       "      <th>yegvx</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>T</td>\n",
       "      <td>0.133333</td>\n",
       "      <td>0.533333</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.066667</td>\n",
       "      <td>0.533333</td>\n",
       "      <td>0.866667</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>0.533333</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.533333</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.533333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>I</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.800000</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.466667</td>\n",
       "      <td>0.133333</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.266667</td>\n",
       "      <td>0.866667</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.600000</td>\n",
       "      <td>0.133333</td>\n",
       "      <td>0.533333</td>\n",
       "      <td>0.266667</td>\n",
       "      <td>0.666667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>D</td>\n",
       "      <td>0.266667</td>\n",
       "      <td>0.733333</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.533333</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.133333</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.466667</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.466667</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.600000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>N</td>\n",
       "      <td>0.466667</td>\n",
       "      <td>0.733333</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.600000</td>\n",
       "      <td>0.266667</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.266667</td>\n",
       "      <td>0.266667</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>0.133333</td>\n",
       "      <td>0.533333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>G</td>\n",
       "      <td>0.133333</td>\n",
       "      <td>0.066667</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.066667</td>\n",
       "      <td>0.066667</td>\n",
       "      <td>0.533333</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.600000</td>\n",
       "      <td>0.066667</td>\n",
       "      <td>0.466667</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.666667</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",
       "      <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>19995</th>\n",
       "      <td>D</td>\n",
       "      <td>0.133333</td>\n",
       "      <td>0.133333</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.133333</td>\n",
       "      <td>0.466667</td>\n",
       "      <td>0.466667</td>\n",
       "      <td>0.466667</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.266667</td>\n",
       "      <td>0.133333</td>\n",
       "      <td>0.533333</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.466667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19996</th>\n",
       "      <td>C</td>\n",
       "      <td>0.466667</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>0.533333</td>\n",
       "      <td>0.533333</td>\n",
       "      <td>0.266667</td>\n",
       "      <td>0.266667</td>\n",
       "      <td>0.533333</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.600000</td>\n",
       "      <td>0.800000</td>\n",
       "      <td>0.600000</td>\n",
       "      <td>0.866667</td>\n",
       "      <td>0.133333</td>\n",
       "      <td>0.600000</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.466667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19997</th>\n",
       "      <td>T</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.600000</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.466667</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.733333</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.466667</td>\n",
       "      <td>0.733333</td>\n",
       "      <td>0.600000</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.133333</td>\n",
       "      <td>0.800000</td>\n",
       "      <td>0.133333</td>\n",
       "      <td>0.266667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19998</th>\n",
       "      <td>S</td>\n",
       "      <td>0.133333</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.266667</td>\n",
       "      <td>0.133333</td>\n",
       "      <td>0.066667</td>\n",
       "      <td>0.533333</td>\n",
       "      <td>0.466667</td>\n",
       "      <td>0.133333</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.533333</td>\n",
       "      <td>0.066667</td>\n",
       "      <td>0.600000</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.533333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19999</th>\n",
       "      <td>A</td>\n",
       "      <td>0.266667</td>\n",
       "      <td>0.600000</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.133333</td>\n",
       "      <td>0.600000</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.066667</td>\n",
       "      <td>0.533333</td>\n",
       "      <td>0.066667</td>\n",
       "      <td>0.533333</td>\n",
       "      <td>0.133333</td>\n",
       "      <td>0.466667</td>\n",
       "      <td>0.133333</td>\n",
       "      <td>0.533333</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>20000 rows × 17 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      letter     x-box     y-box     width      high     onpix     x-bar  \\\n",
       "0          T  0.133333  0.533333  0.200000  0.333333  0.066667  0.533333   \n",
       "1          I  0.333333  0.800000  0.200000  0.466667  0.133333  0.666667   \n",
       "2          D  0.266667  0.733333  0.400000  0.533333  0.400000  0.666667   \n",
       "3          N  0.466667  0.733333  0.400000  0.400000  0.200000  0.333333   \n",
       "4          G  0.133333  0.066667  0.200000  0.066667  0.066667  0.533333   \n",
       "...      ...       ...       ...       ...       ...       ...       ...   \n",
       "19995      D  0.133333  0.133333  0.200000  0.200000  0.133333  0.466667   \n",
       "19996      C  0.466667  0.666667  0.533333  0.533333  0.266667  0.266667   \n",
       "19997      T  0.400000  0.600000  0.400000  0.466667  0.333333  0.400000   \n",
       "19998      S  0.133333  0.200000  0.266667  0.133333  0.066667  0.533333   \n",
       "19999      A  0.266667  0.600000  0.400000  0.400000  0.133333  0.600000   \n",
       "\n",
       "          y-bar     x2bar     y2bar     xybar     x2ybr     xy2br     x-ege  \\\n",
       "0      0.866667  0.000000  0.400000  0.400000  0.666667  0.533333  0.000000   \n",
       "1      0.333333  0.333333  0.266667  0.866667  0.200000  0.600000  0.133333   \n",
       "2      0.400000  0.133333  0.400000  0.666667  0.200000  0.466667  0.200000   \n",
       "3      0.600000  0.266667  0.400000  0.266667  0.266667  0.666667  0.400000   \n",
       "4      0.400000  0.400000  0.400000  0.400000  0.333333  0.600000  0.066667   \n",
       "...         ...       ...       ...       ...       ...       ...       ...   \n",
       "19995  0.466667  0.466667  0.400000  0.400000  0.400000  0.266667  0.133333   \n",
       "19996  0.533333  0.400000  0.600000  0.800000  0.600000  0.866667  0.133333   \n",
       "19997  0.733333  0.200000  0.466667  0.733333  0.600000  0.333333  0.133333   \n",
       "19998  0.466667  0.133333  0.400000  0.666667  0.400000  0.533333  0.066667   \n",
       "19999  0.333333  0.200000  0.066667  0.533333  0.066667  0.533333  0.133333   \n",
       "\n",
       "          xegvy     y-ege     yegvx  \n",
       "0      0.533333  0.000000  0.533333  \n",
       "1      0.533333  0.266667  0.666667  \n",
       "2      0.466667  0.200000  0.600000  \n",
       "3      0.666667  0.133333  0.533333  \n",
       "4      0.466667  0.333333  0.666667  \n",
       "...         ...       ...       ...  \n",
       "19995  0.533333  0.200000  0.466667  \n",
       "19996  0.600000  0.200000  0.466667  \n",
       "19997  0.800000  0.133333  0.266667  \n",
       "19998  0.600000  0.333333  0.533333  \n",
       "19999  0.466667  0.133333  0.533333  \n",
       "\n",
       "[20000 rows x 17 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.columns = ['letter', 'x-box', 'y-box', 'width', 'high', 'onpix', 'x-bar', 'y-bar', 'x2bar', 'y2bar', 'xybar', 'x2ybr', 'xy2br', 'x-ege', 'xegvy', 'y-ege', 'yegvx']\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "504c45ca",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df.drop('letter', axis=1)\n",
    "y = df['letter']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "6f0eb902",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = X.loc[:15999]\n",
    "y_train = y.loc[:15999]\n",
    "X_test = X.loc[16000:]\n",
    "y_test = y.loc[16000:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "5c18a597",
   "metadata": {},
   "outputs": [],
   "source": [
    "def evaluate(model):\n",
    "    model.fit(X_train, y_train)\n",
    "    x_train_prediction = model.predict(X_train)\n",
    "    x_test_prediction = model.predict(X_test)\n",
    "    train_accuracy = accuracy_score(y_train, x_train_prediction)\n",
    "    test_accuracy = accuracy_score(y_test, x_test_prediction)\n",
    "    clf_report = classification_report(y_test, x_test_prediction)\n",
    "    print(f'Train set accuracy: {train_accuracy:.2f}')\n",
    "    print(f'Test set accuracy: {test_accuracy:.2f}' + '\\n')\n",
    "    print(clf_report + '\\n')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "4c85497e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train set accuracy: 0.98\n",
      "Test set accuracy: 0.95\n",
      "\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           A       0.98      0.99      0.98       156\n",
      "           B       0.87      0.97      0.92       136\n",
      "           C       0.96      0.96      0.96       142\n",
      "           D       0.89      0.97      0.93       167\n",
      "           E       0.89      0.94      0.91       152\n",
      "           F       0.92      0.93      0.93       153\n",
      "           G       0.97      0.91      0.94       164\n",
      "           H       0.84      0.87      0.85       151\n",
      "           I       0.93      0.96      0.95       165\n",
      "           J       0.95      0.92      0.93       148\n",
      "           K       0.92      0.90      0.91       146\n",
      "           L       0.99      0.97      0.98       157\n",
      "           M       0.97      0.97      0.97       144\n",
      "           N       0.97      0.90      0.93       166\n",
      "           O       0.91      0.94      0.93       139\n",
      "           P       0.96      0.90      0.93       168\n",
      "           Q       0.95      0.94      0.94       168\n",
      "           R       0.91      0.93      0.92       161\n",
      "           S       0.98      0.95      0.97       161\n",
      "           T       0.99      0.97      0.98       151\n",
      "           U       0.99      0.99      0.99       168\n",
      "           V       0.95      0.99      0.97       136\n",
      "           W       0.99      0.97      0.98       139\n",
      "           X       0.95      0.94      0.94       159\n",
      "           Y       0.98      0.95      0.97       145\n",
      "           Z       0.97      0.97      0.97       158\n",
      "\n",
      "    accuracy                           0.95      4000\n",
      "   macro avg       0.95      0.95      0.95      4000\n",
      "weighted avg       0.95      0.95      0.95      4000\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "model = KNeighborsClassifier()\n",
    "evaluate(model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "58893068",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 10 folds for each of 8 candidates, totalling 80 fits\n",
      "[CV] END ......................................n_neighbors=1; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=1; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=1; total time=   0.2s\n",
      "[CV] END ......................................n_neighbors=1; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=1; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=1; total time=   0.2s\n",
      "[CV] END ......................................n_neighbors=1; total time=   0.2s\n",
      "[CV] END ......................................n_neighbors=1; total time=   0.2s\n",
      "[CV] END ......................................n_neighbors=1; total time=   0.2s\n",
      "[CV] END ......................................n_neighbors=1; total time=   0.2s\n",
      "[CV] END ......................................n_neighbors=2; total time=   0.2s\n",
      "[CV] END ......................................n_neighbors=2; total time=   0.2s\n",
      "[CV] END ......................................n_neighbors=2; total time=   0.2s\n",
      "[CV] END ......................................n_neighbors=2; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=2; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=2; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=2; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=2; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=2; total time=   0.2s\n",
      "[CV] END ......................................n_neighbors=2; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=3; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=3; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=3; total time=   0.2s\n",
      "[CV] END ......................................n_neighbors=3; total time=   0.2s\n",
      "[CV] END ......................................n_neighbors=3; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=3; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=3; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=3; total time=   0.2s\n",
      "[CV] END ......................................n_neighbors=3; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=3; total time=   0.4s\n",
      "[CV] END ......................................n_neighbors=4; total time=   0.4s\n",
      "[CV] END ......................................n_neighbors=4; total time=   0.4s\n",
      "[CV] END ......................................n_neighbors=4; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=4; total time=   0.4s\n",
      "[CV] END ......................................n_neighbors=4; total time=   0.4s\n",
      "[CV] END ......................................n_neighbors=4; total time=   0.5s\n",
      "[CV] END ......................................n_neighbors=4; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=4; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=4; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=4; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=5; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=5; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=5; total time=   0.4s\n",
      "[CV] END ......................................n_neighbors=5; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=5; total time=   0.4s\n",
      "[CV] END ......................................n_neighbors=5; total time=   0.4s\n",
      "[CV] END ......................................n_neighbors=5; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=5; total time=   0.4s\n",
      "[CV] END ......................................n_neighbors=5; total time=   0.4s\n",
      "[CV] END ......................................n_neighbors=5; total time=   0.5s\n",
      "[CV] END ......................................n_neighbors=6; total time=   0.4s\n",
      "[CV] END ......................................n_neighbors=6; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=6; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=6; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=6; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=6; total time=   0.4s\n",
      "[CV] END ......................................n_neighbors=6; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=6; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=6; total time=   0.4s\n",
      "[CV] END ......................................n_neighbors=6; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=7; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=7; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=7; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=7; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=7; total time=   0.4s\n",
      "[CV] END ......................................n_neighbors=7; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=7; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=7; total time=   0.4s\n",
      "[CV] END ......................................n_neighbors=7; total time=   0.5s\n",
      "[CV] END ......................................n_neighbors=7; total time=   0.7s\n",
      "[CV] END ......................................n_neighbors=8; total time=   0.4s\n",
      "[CV] END ......................................n_neighbors=8; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=8; total time=   0.4s\n",
      "[CV] END ......................................n_neighbors=8; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=8; total time=   0.8s\n",
      "[CV] END ......................................n_neighbors=8; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=8; total time=   0.3s\n",
      "[CV] END ......................................n_neighbors=8; total time=   0.4s\n",
      "[CV] END ......................................n_neighbors=8; total time=   0.4s\n",
      "[CV] END ......................................n_neighbors=8; total time=   0.4s\n"
     ]
    }
   ],
   "source": [
    "k_range = list(range(1, 9))\n",
    "param_grid = dict(n_neighbors=k_range)\n",
    "  \n",
    "grid = GridSearchCV(model, param_grid, cv=10, scoring='accuracy', return_train_score=False, verbose=2)\n",
    "  \n",
    "grid_search=grid.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "47340434",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'n_neighbors': 1}\n"
     ]
    }
   ],
   "source": [
    "print(grid_search.best_params_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "b8ca6792",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train set accuracy: 1.00\n",
      "Test set accuracy: 0.96\n",
      "\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           A       0.99      0.99      0.99       156\n",
      "           B       0.91      0.95      0.93       136\n",
      "           C       0.96      0.96      0.96       142\n",
      "           D       0.94      0.97      0.96       167\n",
      "           E       0.95      0.91      0.93       152\n",
      "           F       0.93      0.92      0.92       153\n",
      "           G       0.96      0.96      0.96       164\n",
      "           H       0.91      0.88      0.90       151\n",
      "           I       0.95      0.97      0.96       165\n",
      "           J       0.97      0.94      0.96       148\n",
      "           K       0.91      0.90      0.91       146\n",
      "           L       0.97      0.97      0.97       157\n",
      "           M       0.99      0.98      0.98       144\n",
      "           N       0.97      0.95      0.96       166\n",
      "           O       0.94      0.97      0.95       139\n",
      "           P       0.95      0.93      0.94       168\n",
      "           Q       0.95      0.96      0.95       168\n",
      "           R       0.90      0.91      0.91       161\n",
      "           S       0.98      0.98      0.98       161\n",
      "           T       0.99      0.97      0.98       151\n",
      "           U       0.98      0.99      0.99       168\n",
      "           V       0.94      0.99      0.96       136\n",
      "           W       0.99      0.98      0.98       139\n",
      "           X       0.97      0.96      0.96       159\n",
      "           Y       0.99      0.97      0.98       145\n",
      "           Z       0.96      0.98      0.97       158\n",
      "\n",
      "    accuracy                           0.96      4000\n",
      "   macro avg       0.96      0.96      0.96      4000\n",
      "weighted avg       0.96      0.96      0.96      4000\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "model = KNeighborsClassifier(**grid_search.best_params_)\n",
    "evaluate(model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "e0de36d2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "no_neighbors = np.arange(1, 10, 1)\n",
    "train_accuracy = np.empty(len(no_neighbors))\n",
    "test_accuracy = np.empty(len(no_neighbors))\n",
    "\n",
    "for i, k in enumerate(no_neighbors):\n",
    "    model = KNeighborsClassifier(n_neighbors=k)\n",
    "    model.fit(X_train,y_train)\n",
    "    train_accuracy[i] = model.score(X_train, y_train)\n",
    "    test_accuracy[i] = model.score(X_test, y_test)\n",
    "\n",
    "plt.title('k-nearest neighbors accuracy')\n",
    "plt.plot(no_neighbors, test_accuracy, label='test')\n",
    "plt.plot(no_neighbors, train_accuracy, label='train')\n",
    "plt.legend()\n",
    "plt.xlabel('number of neighbors')\n",
    "plt.ylabel('accuracy')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "6b5476bb",
   "metadata": {},
   "outputs": [],
   "source": [
    "encoder = LabelEncoder()\n",
    "\n",
    "encoder.fit(y_train)\n",
    "train_labels = encoder.transform(y_train)\n",
    "\n",
    "encoder.fit(y_test)\n",
    "test_labels = encoder.transform(y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "9a233b0f",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "1000/1000 [==============================] - 1s 1ms/step - loss: 2.1995 - accuracy: 0.3871 - val_loss: 1.5618 - val_accuracy: 0.5575\n",
      "Epoch 2/200\n",
      "1000/1000 [==============================] - 1s 1ms/step - loss: 1.3519 - accuracy: 0.6142 - val_loss: 1.2612 - val_accuracy: 0.6428\n",
      "Epoch 3/200\n",
      "1000/1000 [==============================] - 1s 1ms/step - loss: 1.1554 - accuracy: 0.6767 - val_loss: 1.1243 - val_accuracy: 0.6985\n",
      "Epoch 4/200\n",
      "1000/1000 [==============================] - 1s 929us/step - loss: 1.0387 - accuracy: 0.7136 - val_loss: 1.0528 - val_accuracy: 0.7045\n",
      "Epoch 5/200\n",
      "1000/1000 [==============================] - 1s 906us/step - loss: 0.9603 - accuracy: 0.7312 - val_loss: 0.9679 - val_accuracy: 0.7337\n",
      "Epoch 6/200\n",
      "1000/1000 [==============================] - 1s 845us/step - loss: 0.9046 - accuracy: 0.7458 - val_loss: 0.9442 - val_accuracy: 0.7247\n",
      "Epoch 7/200\n",
      "1000/1000 [==============================] - 1s 828us/step - loss: 0.8553 - accuracy: 0.7609 - val_loss: 0.9050 - val_accuracy: 0.7320\n",
      "Epoch 8/200\n",
      "1000/1000 [==============================] - 1s 864us/step - loss: 0.8086 - accuracy: 0.7704 - val_loss: 0.8285 - val_accuracy: 0.7660\n",
      "Epoch 9/200\n",
      "1000/1000 [==============================] - 1s 824us/step - loss: 0.7705 - accuracy: 0.7801 - val_loss: 0.7972 - val_accuracy: 0.7770\n",
      "Epoch 10/200\n",
      "1000/1000 [==============================] - 1s 842us/step - loss: 0.7291 - accuracy: 0.7914 - val_loss: 0.7544 - val_accuracy: 0.7820\n",
      "Epoch 11/200\n",
      "1000/1000 [==============================] - 1s 847us/step - loss: 0.6971 - accuracy: 0.7987 - val_loss: 0.7307 - val_accuracy: 0.7815\n",
      "Epoch 12/200\n",
      "1000/1000 [==============================] - 1s 832us/step - loss: 0.6638 - accuracy: 0.8067 - val_loss: 0.7080 - val_accuracy: 0.7928\n",
      "Epoch 13/200\n",
      "1000/1000 [==============================] - 1s 859us/step - loss: 0.6342 - accuracy: 0.8132 - val_loss: 0.6748 - val_accuracy: 0.8000\n",
      "Epoch 14/200\n",
      "1000/1000 [==============================] - 1s 874us/step - loss: 0.6066 - accuracy: 0.8218 - val_loss: 0.6513 - val_accuracy: 0.8045\n",
      "Epoch 15/200\n",
      "1000/1000 [==============================] - 1s 820us/step - loss: 0.5842 - accuracy: 0.8269 - val_loss: 0.6238 - val_accuracy: 0.8095\n",
      "Epoch 16/200\n",
      "1000/1000 [==============================] - 1s 835us/step - loss: 0.5576 - accuracy: 0.8316 - val_loss: 0.6039 - val_accuracy: 0.8117\n",
      "Epoch 17/200\n",
      "1000/1000 [==============================] - 1s 839us/step - loss: 0.5388 - accuracy: 0.8392 - val_loss: 0.5753 - val_accuracy: 0.8285\n",
      "Epoch 18/200\n",
      "1000/1000 [==============================] - 1s 901us/step - loss: 0.5167 - accuracy: 0.8447 - val_loss: 0.5744 - val_accuracy: 0.8217\n",
      "Epoch 19/200\n",
      "1000/1000 [==============================] - 1s 869us/step - loss: 0.4965 - accuracy: 0.8521 - val_loss: 0.5424 - val_accuracy: 0.8375\n",
      "Epoch 20/200\n",
      "1000/1000 [==============================] - 1s 814us/step - loss: 0.4795 - accuracy: 0.8570 - val_loss: 0.5303 - val_accuracy: 0.8405\n",
      "Epoch 21/200\n",
      "1000/1000 [==============================] - 1s 851us/step - loss: 0.4668 - accuracy: 0.8605 - val_loss: 0.5243 - val_accuracy: 0.8443\n",
      "Epoch 22/200\n",
      "1000/1000 [==============================] - 1s 822us/step - loss: 0.4471 - accuracy: 0.8649 - val_loss: 0.5220 - val_accuracy: 0.8415\n",
      "Epoch 23/200\n",
      "1000/1000 [==============================] - 1s 847us/step - loss: 0.4356 - accuracy: 0.8684 - val_loss: 0.4737 - val_accuracy: 0.8550\n",
      "Epoch 24/200\n",
      "1000/1000 [==============================] - 1s 851us/step - loss: 0.4227 - accuracy: 0.8728 - val_loss: 0.4777 - val_accuracy: 0.8575\n",
      "Epoch 25/200\n",
      "1000/1000 [==============================] - 1s 900us/step - loss: 0.4110 - accuracy: 0.8752 - val_loss: 0.4624 - val_accuracy: 0.8640\n",
      "Epoch 26/200\n",
      "1000/1000 [==============================] - 1s 833us/step - loss: 0.4011 - accuracy: 0.8800 - val_loss: 0.4745 - val_accuracy: 0.8540\n",
      "Epoch 27/200\n",
      "1000/1000 [==============================] - 1s 819us/step - loss: 0.3895 - accuracy: 0.8811 - val_loss: 0.4321 - val_accuracy: 0.8690\n",
      "Epoch 28/200\n",
      "1000/1000 [==============================] - 1s 825us/step - loss: 0.3812 - accuracy: 0.8850 - val_loss: 0.4467 - val_accuracy: 0.8662\n",
      "Epoch 29/200\n",
      "1000/1000 [==============================] - 1s 832us/step - loss: 0.3692 - accuracy: 0.8873 - val_loss: 0.4376 - val_accuracy: 0.8640\n",
      "Epoch 30/200\n",
      "1000/1000 [==============================] - 1s 840us/step - loss: 0.3606 - accuracy: 0.8900 - val_loss: 0.4272 - val_accuracy: 0.8695\n",
      "Epoch 31/200\n",
      "1000/1000 [==============================] - 1s 843us/step - loss: 0.3542 - accuracy: 0.8905 - val_loss: 0.4065 - val_accuracy: 0.8765\n",
      "Epoch 32/200\n",
      "1000/1000 [==============================] - 1s 843us/step - loss: 0.3454 - accuracy: 0.8942 - val_loss: 0.4188 - val_accuracy: 0.8662\n",
      "Epoch 33/200\n",
      "1000/1000 [==============================] - 1s 851us/step - loss: 0.3377 - accuracy: 0.8968 - val_loss: 0.3800 - val_accuracy: 0.8838\n",
      "Epoch 34/200\n",
      "1000/1000 [==============================] - 1s 830us/step - loss: 0.3301 - accuracy: 0.8976 - val_loss: 0.3721 - val_accuracy: 0.8865\n",
      "Epoch 35/200\n",
      "1000/1000 [==============================] - 1s 853us/step - loss: 0.3215 - accuracy: 0.9008 - val_loss: 0.3739 - val_accuracy: 0.8845\n",
      "Epoch 36/200\n",
      "1000/1000 [==============================] - 1s 849us/step - loss: 0.3185 - accuracy: 0.8995 - val_loss: 0.3746 - val_accuracy: 0.8845\n",
      "Epoch 37/200\n",
      "1000/1000 [==============================] - 1s 840us/step - loss: 0.3101 - accuracy: 0.9041 - val_loss: 0.3518 - val_accuracy: 0.8905\n",
      "Epoch 38/200\n",
      "1000/1000 [==============================] - 1s 853us/step - loss: 0.3026 - accuracy: 0.9048 - val_loss: 0.3472 - val_accuracy: 0.8910\n",
      "Epoch 39/200\n",
      "1000/1000 [==============================] - 1s 841us/step - loss: 0.2964 - accuracy: 0.9067 - val_loss: 0.3725 - val_accuracy: 0.8838\n",
      "Epoch 40/200\n",
      "1000/1000 [==============================] - 1s 850us/step - loss: 0.2945 - accuracy: 0.9088 - val_loss: 0.3531 - val_accuracy: 0.8863\n",
      "Epoch 41/200\n",
      "1000/1000 [==============================] - 1s 829us/step - loss: 0.2874 - accuracy: 0.9097 - val_loss: 0.3353 - val_accuracy: 0.8940\n",
      "Epoch 42/200\n",
      "1000/1000 [==============================] - 1s 843us/step - loss: 0.2799 - accuracy: 0.9121 - val_loss: 0.3392 - val_accuracy: 0.8923\n",
      "Epoch 43/200\n",
      "1000/1000 [==============================] - 1s 853us/step - loss: 0.2731 - accuracy: 0.9144 - val_loss: 0.3393 - val_accuracy: 0.8947\n",
      "Epoch 44/200\n",
      "1000/1000 [==============================] - 1s 848us/step - loss: 0.2695 - accuracy: 0.9140 - val_loss: 0.3265 - val_accuracy: 0.8972\n",
      "Epoch 45/200\n",
      "1000/1000 [==============================] - 1s 856us/step - loss: 0.2660 - accuracy: 0.9153 - val_loss: 0.3243 - val_accuracy: 0.8972\n",
      "Epoch 46/200\n",
      "1000/1000 [==============================] - 1s 848us/step - loss: 0.2603 - accuracy: 0.9186 - val_loss: 0.3107 - val_accuracy: 0.9053\n",
      "Epoch 47/200\n",
      "1000/1000 [==============================] - 1s 849us/step - loss: 0.2570 - accuracy: 0.9192 - val_loss: 0.3083 - val_accuracy: 0.9025\n",
      "Epoch 48/200\n",
      "1000/1000 [==============================] - 1s 839us/step - loss: 0.2521 - accuracy: 0.9208 - val_loss: 0.3155 - val_accuracy: 0.8992\n",
      "Epoch 49/200\n",
      "1000/1000 [==============================] - 1s 856us/step - loss: 0.2522 - accuracy: 0.9201 - val_loss: 0.3067 - val_accuracy: 0.9022\n",
      "Epoch 50/200\n",
      "1000/1000 [==============================] - 1s 854us/step - loss: 0.2425 - accuracy: 0.9228 - val_loss: 0.2833 - val_accuracy: 0.9087\n",
      "Epoch 51/200\n",
      "1000/1000 [==============================] - 1s 843us/step - loss: 0.2395 - accuracy: 0.9254 - val_loss: 0.2960 - val_accuracy: 0.9035\n",
      "Epoch 52/200\n",
      "1000/1000 [==============================] - 1s 885us/step - loss: 0.2368 - accuracy: 0.9262 - val_loss: 0.2980 - val_accuracy: 0.9053\n",
      "Epoch 53/200\n",
      "1000/1000 [==============================] - 1s 905us/step - loss: 0.2302 - accuracy: 0.9287 - val_loss: 0.2870 - val_accuracy: 0.9115\n",
      "Epoch 54/200\n",
      "1000/1000 [==============================] - 1s 958us/step - loss: 0.2303 - accuracy: 0.9271 - val_loss: 0.2913 - val_accuracy: 0.9093\n",
      "Epoch 55/200\n",
      "1000/1000 [==============================] - 1s 910us/step - loss: 0.2218 - accuracy: 0.9283 - val_loss: 0.2892 - val_accuracy: 0.9065\n",
      "Epoch 56/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1000/1000 [==============================] - 1s 905us/step - loss: 0.2245 - accuracy: 0.9293 - val_loss: 0.2947 - val_accuracy: 0.9032\n",
      "Epoch 57/200\n",
      "1000/1000 [==============================] - 1s 1ms/step - loss: 0.2187 - accuracy: 0.9304 - val_loss: 0.2675 - val_accuracy: 0.9135\n",
      "Epoch 58/200\n",
      "1000/1000 [==============================] - 1s 799us/step - loss: 0.2161 - accuracy: 0.9311 - val_loss: 0.2841 - val_accuracy: 0.9093\n",
      "Epoch 59/200\n",
      "1000/1000 [==============================] - 1s 1ms/step - loss: 0.2132 - accuracy: 0.9311 - val_loss: 0.2657 - val_accuracy: 0.9160\n",
      "Epoch 60/200\n",
      "1000/1000 [==============================] - 1s 956us/step - loss: 0.2096 - accuracy: 0.9331 - val_loss: 0.2812 - val_accuracy: 0.9070\n",
      "Epoch 61/200\n",
      "1000/1000 [==============================] - 1s 799us/step - loss: 0.2106 - accuracy: 0.9318 - val_loss: 0.2592 - val_accuracy: 0.9200\n",
      "Epoch 62/200\n",
      "1000/1000 [==============================] - 1s 787us/step - loss: 0.2040 - accuracy: 0.9343 - val_loss: 0.2652 - val_accuracy: 0.9145\n",
      "Epoch 63/200\n",
      "1000/1000 [==============================] - 1s 790us/step - loss: 0.2015 - accuracy: 0.9372 - val_loss: 0.2634 - val_accuracy: 0.9158\n",
      "Epoch 64/200\n",
      "1000/1000 [==============================] - 1s 787us/step - loss: 0.1993 - accuracy: 0.9358 - val_loss: 0.2683 - val_accuracy: 0.9153\n",
      "Epoch 65/200\n",
      "1000/1000 [==============================] - 1s 795us/step - loss: 0.1954 - accuracy: 0.9364 - val_loss: 0.2571 - val_accuracy: 0.9160\n",
      "Epoch 66/200\n",
      "1000/1000 [==============================] - 1s 794us/step - loss: 0.1939 - accuracy: 0.9361 - val_loss: 0.2692 - val_accuracy: 0.9145\n",
      "Epoch 67/200\n",
      "1000/1000 [==============================] - 1s 938us/step - loss: 0.1902 - accuracy: 0.9376 - val_loss: 0.2711 - val_accuracy: 0.9190\n",
      "Epoch 68/200\n",
      "1000/1000 [==============================] - 1s 805us/step - loss: 0.1871 - accuracy: 0.9395 - val_loss: 0.2508 - val_accuracy: 0.9230\n",
      "Epoch 69/200\n",
      "1000/1000 [==============================] - 1s 790us/step - loss: 0.1853 - accuracy: 0.9405 - val_loss: 0.2614 - val_accuracy: 0.9145\n",
      "Epoch 70/200\n",
      "1000/1000 [==============================] - 1s 816us/step - loss: 0.1841 - accuracy: 0.9407 - val_loss: 0.2565 - val_accuracy: 0.9170\n",
      "Epoch 71/200\n",
      "1000/1000 [==============================] - 1s 799us/step - loss: 0.1785 - accuracy: 0.9439 - val_loss: 0.2512 - val_accuracy: 0.9153\n",
      "Epoch 72/200\n",
      "1000/1000 [==============================] - 1s 957us/step - loss: 0.1795 - accuracy: 0.9431 - val_loss: 0.2719 - val_accuracy: 0.9135\n",
      "Epoch 73/200\n",
      "1000/1000 [==============================] - 1s 849us/step - loss: 0.1764 - accuracy: 0.9426 - val_loss: 0.2457 - val_accuracy: 0.9220\n",
      "Epoch 74/200\n",
      "1000/1000 [==============================] - 1s 1ms/step - loss: 0.1750 - accuracy: 0.9442 - val_loss: 0.2578 - val_accuracy: 0.9158\n",
      "Epoch 75/200\n",
      "1000/1000 [==============================] - 1s 1ms/step - loss: 0.1712 - accuracy: 0.9436 - val_loss: 0.2431 - val_accuracy: 0.9225\n",
      "Epoch 76/200\n",
      "1000/1000 [==============================] - 1s 795us/step - loss: 0.1708 - accuracy: 0.9437 - val_loss: 0.2409 - val_accuracy: 0.9215\n",
      "Epoch 77/200\n",
      "1000/1000 [==============================] - 1s 802us/step - loss: 0.1682 - accuracy: 0.9448 - val_loss: 0.2325 - val_accuracy: 0.9260\n",
      "Epoch 78/200\n",
      "1000/1000 [==============================] - 1s 1ms/step - loss: 0.1662 - accuracy: 0.9456 - val_loss: 0.2302 - val_accuracy: 0.9275\n",
      "Epoch 79/200\n",
      "1000/1000 [==============================] - 1s 1ms/step - loss: 0.1647 - accuracy: 0.9467 - val_loss: 0.2231 - val_accuracy: 0.9275\n",
      "Epoch 80/200\n",
      "1000/1000 [==============================] - 1s 1ms/step - loss: 0.1632 - accuracy: 0.9466 - val_loss: 0.2489 - val_accuracy: 0.9230\n",
      "Epoch 81/200\n",
      "1000/1000 [==============================] - 1s 887us/step - loss: 0.1600 - accuracy: 0.9466 - val_loss: 0.2392 - val_accuracy: 0.9212\n",
      "Epoch 82/200\n",
      "1000/1000 [==============================] - 1s 834us/step - loss: 0.1612 - accuracy: 0.9466 - val_loss: 0.2222 - val_accuracy: 0.9285\n",
      "Epoch 83/200\n",
      "1000/1000 [==============================] - 1s 867us/step - loss: 0.1573 - accuracy: 0.9501 - val_loss: 0.2414 - val_accuracy: 0.9220\n",
      "Epoch 84/200\n",
      "1000/1000 [==============================] - 1s 885us/step - loss: 0.1553 - accuracy: 0.9486 - val_loss: 0.2256 - val_accuracy: 0.9290\n",
      "Epoch 85/200\n",
      "1000/1000 [==============================] - 1s 908us/step - loss: 0.1538 - accuracy: 0.9488 - val_loss: 0.2414 - val_accuracy: 0.9222\n",
      "Epoch 86/200\n",
      "1000/1000 [==============================] - 1s 849us/step - loss: 0.1534 - accuracy: 0.9493 - val_loss: 0.2210 - val_accuracy: 0.9298\n",
      "Epoch 87/200\n",
      "1000/1000 [==============================] - 1s 836us/step - loss: 0.1507 - accuracy: 0.9507 - val_loss: 0.2396 - val_accuracy: 0.9205\n",
      "Epoch 88/200\n",
      "1000/1000 [==============================] - 1s 814us/step - loss: 0.1490 - accuracy: 0.9504 - val_loss: 0.2245 - val_accuracy: 0.9285\n",
      "Epoch 89/200\n",
      "1000/1000 [==============================] - 1s 833us/step - loss: 0.1486 - accuracy: 0.9516 - val_loss: 0.2263 - val_accuracy: 0.9262\n",
      "Epoch 90/200\n",
      "1000/1000 [==============================] - 1s 836us/step - loss: 0.1468 - accuracy: 0.9506 - val_loss: 0.2248 - val_accuracy: 0.9270\n",
      "Epoch 91/200\n",
      "1000/1000 [==============================] - 1s 834us/step - loss: 0.1454 - accuracy: 0.9513 - val_loss: 0.2253 - val_accuracy: 0.9300\n",
      "Epoch 92/200\n",
      "1000/1000 [==============================] - 1s 827us/step - loss: 0.1416 - accuracy: 0.9523 - val_loss: 0.2207 - val_accuracy: 0.9293\n",
      "Epoch 93/200\n",
      "1000/1000 [==============================] - 1s 839us/step - loss: 0.1402 - accuracy: 0.9548 - val_loss: 0.2352 - val_accuracy: 0.9265\n",
      "Epoch 94/200\n",
      "1000/1000 [==============================] - 1s 872us/step - loss: 0.1391 - accuracy: 0.9539 - val_loss: 0.2237 - val_accuracy: 0.9298\n",
      "Epoch 95/200\n",
      "1000/1000 [==============================] - 1s 830us/step - loss: 0.1397 - accuracy: 0.9526 - val_loss: 0.2019 - val_accuracy: 0.9352\n",
      "Epoch 96/200\n",
      "1000/1000 [==============================] - 1s 837us/step - loss: 0.1366 - accuracy: 0.9539 - val_loss: 0.2237 - val_accuracy: 0.9283\n",
      "Epoch 97/200\n",
      "1000/1000 [==============================] - 1s 832us/step - loss: 0.1336 - accuracy: 0.9567 - val_loss: 0.2263 - val_accuracy: 0.9252\n",
      "Epoch 98/200\n",
      "1000/1000 [==============================] - 1s 839us/step - loss: 0.1326 - accuracy: 0.9567 - val_loss: 0.2215 - val_accuracy: 0.9277\n",
      "Epoch 99/200\n",
      "1000/1000 [==============================] - 1s 840us/step - loss: 0.1312 - accuracy: 0.9559 - val_loss: 0.2199 - val_accuracy: 0.9280\n",
      "Epoch 100/200\n",
      "1000/1000 [==============================] - 1s 825us/step - loss: 0.1314 - accuracy: 0.9554 - val_loss: 0.2539 - val_accuracy: 0.9247\n",
      "Epoch 101/200\n",
      "1000/1000 [==============================] - 1s 829us/step - loss: 0.1328 - accuracy: 0.9549 - val_loss: 0.2324 - val_accuracy: 0.9308\n",
      "Epoch 102/200\n",
      "1000/1000 [==============================] - 1s 850us/step - loss: 0.1304 - accuracy: 0.9579 - val_loss: 0.2238 - val_accuracy: 0.9330\n",
      "Epoch 103/200\n",
      "1000/1000 [==============================] - 1s 818us/step - loss: 0.1274 - accuracy: 0.9576 - val_loss: 0.2231 - val_accuracy: 0.9285\n",
      "Epoch 104/200\n",
      "1000/1000 [==============================] - 1s 826us/step - loss: 0.1291 - accuracy: 0.9565 - val_loss: 0.2256 - val_accuracy: 0.9252\n",
      "Epoch 105/200\n",
      "1000/1000 [==============================] - 1s 820us/step - loss: 0.1249 - accuracy: 0.9571 - val_loss: 0.2157 - val_accuracy: 0.9317\n"
     ]
    }
   ],
   "source": [
    "model = Sequential()\n",
    "model.add(Dense(units=64, activation='relu', input_shape=(16,)))\n",
    "model.add(Dense(units=64, activation='relu'))\n",
    "model.add(Dense(units=26, activation='softmax'))\n",
    "\n",
    "model.compile(loss='sparse_categorical_crossentropy',\n",
    "              optimizer='adam',\n",
    "              metrics=['accuracy'])\n",
    "\n",
    "early_stopping = callbacks.EarlyStopping(\n",
    "    monitor='val_loss',\n",
    "    min_delta=0.001, \n",
    "    patience=10, \n",
    "    restore_best_weights=True)\n",
    "\n",
    "history = model.fit(X_train, train_labels, validation_data=(X_test, test_labels), epochs=200, batch_size=16, callbacks=[early_stopping])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "d3a75ca6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train set accuracy: 0.96\n"
     ]
    }
   ],
   "source": [
    "x_train_prediction = model.predict(X_train) \n",
    "y_classes = x_train_prediction.argmax(axis=-1)\n",
    "train_accuracy = accuracy_score(train_labels, y_classes)\n",
    "print(f'Train set accuracy: {train_accuracy:.2f}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "5f5bbb8c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test set accuracy: 0.94\n"
     ]
    }
   ],
   "source": [
    "x_test_prediction = model.predict(X_test) \n",
    "y_classes = x_test_prediction.argmax(axis=-1)\n",
    "test_accuracy = accuracy_score(test_labels, y_classes)\n",
    "print(f'Test set accuracy: {test_accuracy:.2f}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "4d409db1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           A       0.94      0.94      0.94       156\n",
      "           B       0.88      0.98      0.93       136\n",
      "           C       0.95      0.93      0.94       142\n",
      "           D       0.88      0.96      0.92       167\n",
      "           E       0.92      0.91      0.91       152\n",
      "           F       0.90      0.95      0.92       153\n",
      "           G       0.88      0.92      0.90       164\n",
      "           H       0.95      0.83      0.88       151\n",
      "           I       0.94      0.92      0.93       165\n",
      "           J       0.92      0.93      0.93       148\n",
      "           K       0.93      0.92      0.93       146\n",
      "           L       0.97      0.97      0.97       157\n",
      "           M       0.95      0.99      0.97       144\n",
      "           N       0.96      0.91      0.93       166\n",
      "           O       0.93      0.88      0.90       139\n",
      "           P       0.96      0.89      0.92       168\n",
      "           Q       0.96      0.91      0.94       168\n",
      "           R       0.87      0.94      0.90       161\n",
      "           S       0.95      0.94      0.95       161\n",
      "           T       0.95      0.97      0.96       151\n",
      "           U       0.99      0.97      0.98       168\n",
      "           V       0.98      0.95      0.96       136\n",
      "           W       0.98      0.96      0.97       139\n",
      "           X       0.94      0.95      0.95       159\n",
      "           Y       0.93      0.95      0.94       145\n",
      "           Z       0.95      0.96      0.95       158\n",
      "\n",
      "    accuracy                           0.94      4000\n",
      "   macro avg       0.94      0.94      0.94      4000\n",
      "weighted avg       0.94      0.94      0.94      4000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "clf_report = classification_report(encoder.inverse_transform(test_labels), encoder.inverse_transform(y_classes))\n",
    "print(clf_report)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "98e0f9ee",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['accuracy'])\n",
    "plt.plot(history.history['val_accuracy'])\n",
    "plt.title('neural network accuracy')\n",
    "plt.ylabel('accuracy')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train', 'val'], loc='upper left')\n",
    "plt.show()\n",
    "\n",
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.title('neural network loss')\n",
    "plt.ylabel('loss')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train', 'val'], loc='upper left')\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}