HeartGuardAI - AI Powered Heart Disease Predictor

📌 Project Goal

HeartGuardAI is designed to assist in the early detection of heart disease by leveraging machine learning models.
The goal is to provide a simple, accurate, and accessible tool for predicting the likelihood of heart disease based on patient data.
The project aims to support preventive healthcare and reduce diagnostic delays.

---

📖 Overview

HeartGuardAI uses a trained deep learning model to predict whether a patient is likely to have heart disease.
It processes key health parameters such as age, cholesterol levels, and blood pressure, then delivers predictions through an easy-to-use Streamlit web interface.
This project demonstrates the complete ML pipeline from data preprocessing to deployment.

---

🔄 Project Workflow

1️⃣ Data Preprocessing

• Loaded heart disease dataset from UCI Machine Learning Repository.
• Checked for missing/null values and handled inconsistencies.
• Encoded categorical variables into numerical format.
• Normalized and scaled features using StandardScaler from scikit-learn.
• Split dataset into 80% training and 20% testing sets for model evaluation.

2️⃣ Model Building

Tested multiple classification models:

• Logistic Regression
• Random Forest Classifier
• Support Vector Machine (SVM)
• Decision Tree Classifier
• K-Nearest Neighbors (KNN)
• Gradient Boosting Classifier
• Deep Neural Network (✅ Best Performer)

Final Model:

• Architecture: 3 Dense layers with ReLU activation and Dropout layers to prevent overfitting
• Output Layer: Sigmoid activation for binary classification
• Optimizer: Adam
• Loss: Binary Crossentropy
• Saved using .keras format for deployment and joblib for the scaler.

3️⃣ Evaluation Metrics

• Accuracy: 0.86 on the test set
• Precision: 0.84
• Recall: 0.87
• F1-Score: 0.85
• Confusion matrix visualized for better understanding of model performance.

4️⃣ Deployment

• Built an interactive Streamlit web app for local deployment.
• User inputs health data through sliders and dropdowns.
• Data is scaled using the saved StandardScaler before prediction.
• Model predicts and displays results instantly with a clear "Heart Disease Detected" or "No Heart Disease Detected" message.
• Ready to be deployed on platforms like Streamlit Cloud or Heroku for global access.

---

🛠 Tech Stack

Programming Language:

• Python

Libraries & Frameworks:

• Data Preprocessing & Analysis: pandas, numpy, scikit-learn
• Visualization: matplotlib, seaborn
• Model Building: tensorflow, keras
• Deployment: streamlit
• Model Persistence: joblib

---

📂 Project Structure

HeartGuardAI/
├── Dataset/                             # Folder containing the dataset
│   └── heart.csv                        # Main dataset used for training the model
├── .gitignore                           # Prevents Git from tracking unnecessary files
├── Heart_Disease_Prediction.ipynb       # Jupyter notebook for model training & evaluation
├── LICENSE                              # Allows reuse, with attribution, no warranty
├── README.md                            # Project documentation
├── app.py                               # Streamlit app for deployment
├── model.keras                          # Saved trained model
├── requirements.txt                     # Python dependencies
└── scaler.pkl                           # Saved StandardScaler for preprocessing

---

✨ Features

• Predicts heart disease likelihood from 13 health parameters.
• Real-time prediction via a simple web interface.
• Fast and lightweight for local execution.
• Visualization of dataset patterns and model performance.

---

✨ Future Enhancements

• Predicts heart disease likelihood from 13 health parameters.
• Real-time prediction via a simple web interface.
• Fast and lightweight for local execution.
• Visualization of dataset patterns and model performance.

---

📜 License

This project is licensed under the MIT License.

---

📬 Contact

---

Thanks for visiting ! Feel free to explore my other repositories and connect with me. 🚀