Heart Disease Prediction Web App

A Streamlit-based web application for predicting the risk of heart disease using machine learning models.

Features

• User-Friendly Interface: Input patient data via sidebar sliders and dropdowns.
• Multiple Models: Choose between SVM, KNN, Logistic Regression, Decision Tree, or a Voting Classifier.
• Visualizations:
  • Accuracy comparison across models.
  • Feature correlation matrix.
  • Decision Tree structure.
  • Confusion matrices for individual models.
• Real-Time Prediction: Instantly assess heart disease risk (low/high) based on input data.

Installation

1. Clone the repository:

   git clone https://github.com/your-username/heart-disease-prediction.git
   cd heart-disease-prediction

2. Install dependencies:

   pip install -r requirements.txt

   (Create requirements.txt with: streamlit, scikit-learn, pandas, numpy, matplotlib, seaborn)

3. Run the app:

   streamlit run model.py

Usage

1. Input patient information in the sidebar (e.g., age, cholesterol, blood pressure).
2. Select a model from the dropdown.
3. Click Predict to see the risk assessment.
4. Explore visualizations for model performance and data insights.

Dataset

The heart.csv file contains clinical features for heart disease prediction, including:

• age, sex, cp (chest pain type), trestbps (resting blood pressure), chol (cholesterol), and more.
• Preprocessing:
  • Missing values and duplicates are removed.
  • Features are scaled using StandardScaler.
  • Lower-priority features (bottom 33% by Z-score) are excluded.

Models

• Supported Algorithms:
  • Support Vector Machine (SVM)
  • K-Nearest Neighbors (KNN)
  • Logistic Regression
  • Decision Tree
  • Voting Classifier (ensemble of all models)
• Metrics:
  • Accuracy scores for each model.
  • Confusion matrices to evaluate performance.

Project Structure

.
├── model.py             # Streamlit app and ML model training
├── heart.csv            # Heart disease dataset
├── README.md            # Project documentation
├── AI-Project.ipynb     # Jupyter Notebook file
└── requirements.txt     # Dependency list