🧠 Detection of Parkinson's Disease ⚕️

Detecting Parkinson's Disease using machine learning and vocal data analysis.

---

🧬 About Parkinson's Disease

Parkinson's Disease is a neurodegenerative disorder that affects movement and worsens over time. It is primarily characterized by tremors, stiffness, and difficulty with balance and coordination. This condition is caused by the gradual breakdown of dopamine-producing neurons in the brain. Early detection of Parkinson's can significantly improve quality of life by allowing for early intervention and better management of symptoms.

Key Symptoms:

• Tremors (shaking)
• Rigidity (stiff muscles)
• Bradykinesia (slowness of movement)
• Impaired balance and coordination

Image Source: Representation of brain areas affected by Parkinson's Disease

🌐 Project Summary

This project leverages Machine Learning to detect Parkinson's Disease by analyzing vocal features. Using a variety of models, we aim to identify patterns in voice data that indicate early symptoms of Parkinson's.

---

📋 Project Workflow

1. Data Collection: Accessed via a reputable source.
2. Data Preprocessing: Cleaned, normalized, and prepared for model training.
3. Exploratory Data Analysis (EDA): Uncovered trends and relationships in vocal features.
4. Balancing & Scaling: Ensured data is balanced and features are appropriately scaled.
5. Model Training & Evaluation: Tested various models to find the most effective one.

---

📊 Dataset Details

• Dataset Used: Parkinson's Disease Dataset
• Source: UCI Machine Learning Repository
• Description: Contains vocal measurements from individuals to help distinguish between Parkinson's and healthy conditions.

🧩 Machine Learning Models

Several algorithms were trained and evaluated to find the best fit for our data:

• Decision Tree Classifier
• Random Forest Classifier
• Logistic Regression
• Support Vector Machine (SVM)
• Naive Bayes
• K-Nearest Neighbors (KNN)
• XGBoost

🎖️ Best Model Performance

The Random Forest Classifier achieved the highest performance:

• Accuracy: 99.6%
• F1 Score: 0.961
• R² Score: 0.862

These metrics reflect the model's effectiveness in detecting Parkinson's indicators from vocal features.

---

📈 Results & Analysis

Metric	Random Forest Classifier
Accuracy	99.6%
F1 Score	0.961
R² Score	0.862

Insights: The Random Forest model showed superior performance due to its capacity to handle complex data patterns, making it a suitable choice for vocal feature classification.

🤝 Contributions

Contributions are welcome! If you’d like to add improvements or suggest enhancements, feel free to create a pull request.

📞 Contact Information

For any inquiries or collaboration opportunities, feel free to reach out:

• Name: Mayank Yadav
• Email: mayanky075@gmail.com
• LinkedIn: LinkedIn Profile
• GitHub: GitHub

I'm always open to discussing new projects, ideas, or opportunities!

---