JenniferAnnKok/Heart_Disease_Classifier

♥ Heart Disease Classifier

An interactive Gradio app that trains a Random Forest classifier on heart disease data and visualises the full ML pipeline — right in your browser.

Features

• 📂 Upload your own CSV — any dataset with the expected columns works
• 📊 Data overview — missing values, dtypes, target distribution
• 🎯 Model metrics — Accuracy, Recall, Precision, F1, OOB score
• 🔀 Confusion Matrix & ROC Curve — visual evaluation
• 🔁 10-Fold Cross Validation — per-fold bar charts + mean ± std
• 🌲 Feature Importance — top 15 Random Forest features

Expected CSV Columns

Column	Type
Age, RestingBP, Cholesterol, FastingBS, MaxHR, Oldpeak	Numeric
Sex, ChestPainType, RestingECG, ExerciseAngina, ST_Slope	Categorical
HeartDisease	Target (0 or 1)

Compatible with the Kaggle Heart Failure Prediction Dataset.

Pipeline

CSV Upload
  └─► Train/Test Split (80/20, stratified)
        └─► OneHotEncoder (categorical cols)
              └─► StandardScaler (numeric cols)
                    └─► RandomForestClassifier(n_estimators=200, class_weight='balanced')
                          └─► Metrics + Visualisations