🌱 Project Story — HealthGuard AI

💡 Inspiration

Healthcare access is still limited in many rural and underserved areas, where people often get diagnosed too late. While learning Machine Learning, I realized that even simple models trained on basic health data could help identify disease risk early. This motivated me to build HealthGuard AI — a low-cost, accessible screening tool for early disease risk prediction.

🧠 What I Learned

This project helped me understand how real-world ML differs from theory. I learned how to:

• Work with medical datasets and clean invalid values
• Handle imbalanced data using SMOTE
• Evaluate models using recall and F1-score instead of accuracy alone
• Build and deploy an end-to-end ML application
• Create a mobile-friendly UI using Streamlit ## 🛠️ How I Built It I cleaned and preprocessed a public medical dataset, engineered meaningful features, and trained multiple ML models. Since missing disease cases is costly in healthcare, I optimized the model for recall. The final model was deployed as a Streamlit web app that allows users to input basic health parameters and receive a disease risk estimate in real time. ## ⚠️ Challenges Some of the key challenges included handling imbalanced data, interpreting medical features correctly, and resolving deployment issues related to dependencies. These challenges taught me valuable lessons about building production-ready ML systems. ## 🌍 Impact HealthGuard AI is designed as an early screening and decision-support tool, not a diagnostic system. It demonstrates how Machine Learning can be used responsibly to improve preventive healthcare access.

Early detection saves lives — and AI can help make it accessible.

Built With

• git
• joblib
• machine-learning
• numpy
• numpy-**imbalanced-data-handling:**-imbalanced-learn-(smote)-**model-persistence:**-joblib-**web-framework:**-streamlit-**visualization:**-matplotlib
• pandas
• python
• random-forest
• seaborn
• streamlit
• svm)-**data-processing:**-pandas