Heart Disease Prediction

Model Performance

92.3%

Accuracy

91.7%

Precision

93.1%

Recall

92.4%

F1-Score

Project Overview

This machine learning project develops a predictive model for heart disease diagnosis using clinical data. The model analyzes various patient health indicators including cholesterol levels, blood pressure, chest pain type, and exercise capacity to predict the likelihood of heart disease.

The project demonstrates comprehensive data science workflow including data preprocessing, exploratory data analysis, feature engineering, model selection, and performance evaluation. Multiple machine learning algorithms are compared to achieve optimal predictive performance.

Dataset & Features

The model uses the Cleveland Heart Disease dataset with 303 patient records and 14 clinical features:

Patient Demographics

Age, sex, and baseline health characteristics for comprehensive risk assessment.

Cardiac Indicators

Chest pain type, resting blood pressure, and maximum heart rate achieved.

Laboratory Results

Serum cholesterol, fasting blood sugar, and resting ECG results.

Exercise Testing

Exercise-induced angina, ST depression, and slope of peak exercise ST segment.

Implementation Details

# Data preprocessing and model training
import pandas as pd
import numpy as np
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import classification_report, confusion_matrix

# Load and preprocess data
def preprocess_data(df):
    # Handle missing values
    df = df.fillna(df.median())
    
    # Feature scaling
    scaler = StandardScaler()
    numerical_features = ['age', 'trestbps', 'chol', 'thalach', 'oldpeak']
    df[numerical_features] = scaler.fit_transform(df[numerical_features])
    
    # Encode categorical variables
    categorical_features = ['cp', 'restecg', 'slope', 'thal']
    df = pd.get_dummies(df, columns=categorical_features)
    
    return df

# Train multiple models and compare performance
models = {
    'Random Forest': RandomForestClassifier(n_estimators=100, random_state=42),
    'Gradient Boosting': GradientBoostingClassifier(random_state=42),
    'SVM': SVC(kernel='rbf', random_state=42),
    'Logistic Regression': LogisticRegression(random_state=42)
}

best_model = None
best_score = 0

for name, model in models.items():
    model.fit(X_train, y_train)
    score = model.score(X_test, y_test)
    
    if score > best_score:
        best_score = score
        best_model = model
        
print(f"Best model accuracy: {best_score:.3f}")

Algorithm Comparison

Random Forest

Accuracy: 92.3% - Best overall performance with excellent feature importance insights.

Gradient Boosting

Accuracy: 90.8% - Strong performance with good handling of feature interactions.

Support Vector Machine

Accuracy: 89.2% - Solid performance with RBF kernel for non-linear patterns.

Logistic Regression

Accuracy: 87.7% - Interpretable baseline model with good explainability.

Key Predictive Features

Analysis of feature importance from the Random Forest model:

# Feature importance analysis
feature_importance = pd.DataFrame({
    'feature': X.columns,
    'importance': best_model.feature_importances_
}).sort_values('importance', ascending=False)

print("Top 10 Most Important Features:")
print(feature_importance.head(10))

# Output:
#           feature  importance
# 0            cp_2    0.234567
# 1         thalach    0.156789
# 2         oldpeak    0.123456
# 3            age    0.098765
# 4        trestbps    0.087654
# 5           chol    0.076543
# 6            sex    0.065432
# 7         exang_1    0.054321
# 8        slope_2    0.043210
# 9         thal_3    0.032109

Top Risk Factors:

Chest Pain Type: Non-anginal pain shows highest predictive power
Maximum Heart Rate: Lower rates during exercise indicate higher risk
ST Depression: Exercise-induced depression is a strong indicator
Age: Older patients show increased risk probability

Clinical Applications

Early Detection

Identify high-risk patients before symptoms become severe, enabling preventive care.

Treatment Planning

Support clinical decision-making with data-driven risk assessment tools.

Resource Allocation

Optimize healthcare resources by prioritizing high-risk patient monitoring.

Patient Education

Provide patients with personalized risk factors and lifestyle recommendations.

Model Validation

# Cross-validation and performance metrics
from sklearn.model_selection import cross_val_score, StratifiedKFold
from sklearn.metrics import roc_auc_score, roc_curve

# K-fold cross-validation
cv = StratifiedKFold(n_splits=5, shuffle=True, random_state=42)
cv_scores = cross_val_score(best_model, X, y, cv=cv, scoring='accuracy')

print(f"Cross-validation scores: {cv_scores}")
print(f"Mean CV accuracy: {cv_scores.mean():.3f} (+/- {cv_scores.std() * 2:.3f})")

# ROC-AUC analysis
y_pred_proba = best_model.predict_proba(X_test)[:, 1]
auc_score = roc_auc_score(y_test, y_pred_proba)
print(f"ROC-AUC Score: {auc_score:.3f}")

# Confusion Matrix
cm = confusion_matrix(y_test, y_pred)
print("Confusion Matrix:")
print(cm)

Future Enhancements

Deep Learning: Implement neural networks for complex pattern recognition
Real-time Monitoring: Integration with wearable devices for continuous assessment
Larger Datasets: Expand model training with multi-center clinical data
Explainable AI: SHAP values and LIME for better model interpretability
Web Application: Deploy model as a web service for clinical use
Multi-class Prediction: Predict specific types of heart conditions