Project information

• Category: Python, AI
• Model Type: Multiple
• Project Date: 25 February, 2025
• Project Link: Github

This project focuses on predicting heart attacks using machine learning, leveraging structured health and lifestyle data to identify individuals at risk. By developing and fine-tuning models such as XGBoost, LightGBM, Logistic Regression, SVM, and Neural Networks, I optimized predictive performance through feature engineering, hyperparameter tuning, and ensemble techniques. The final models were evaluated based on ROC AUC, recall, and precision, ensuring their clinical applicability. With Logistic Regression emerging as the most interpretable and balanced model (76% recall, 20% precision, ROC AUC 0.877), this work highlights the potential of AI-driven risk assessment in preventive healthcare.