dexter6102/hear_failure_detection

🫀 Heart Failure Risk Predictor

A machine learning–powered clinical decision support tool that predicts the risk of death from heart failure using patient clinical data.

Built with Scikit-learn, XGBoost, and SHAP explainability, and deployed via Streamlit.

🚀 Live App

👉 Try it here: https://huggingface.co/spaces/YOURUSERNAME/YOURSPACE_NAME

🧠 What This App Does

• —Predicts probability of mortality in heart failure patients
• —Classifies risk into:
• —✅ Low Risk
• —⚠️ Medium Risk
• —🚨 High Risk
• —Provides SHAP explanations showing why the model made its prediction
• —Displays key clinical insights based on input values

📊 Dataset

• —Source: Heart Failure Clinical Records (Kaggle)
• —Link: https://www.kaggle.com/datasets/andrewmvd/heart-failure-clinical-data
• —Size:
• —299 patients
• —13 clinical features
• —Binary outcome (DEATH_EVENT)
• —Origin: Faisalabad, Pakistan (2015)