AI-Powered Analysis

Cardiovascular Disease Prediction Model

Leveraging advanced machine learning algorithms to analyze health metrics and identify potential cardiovascular risks with high precision.

+2.1%

73.5%

Accuracy

Overall model correctness

+1.4%

75.0%

Precision

Positive predictive value

+0.8%

69.0%

Recall

Sensitivity / True Positive Rate

+1.2%

72.0%

F1 Score

Balance of precision & recall

Model Architecture

We utilize a Gradient Boosting Classifier, chosen for its superior performance on structured tabular data. The model was optimized through extensive hyperparameter tuning using Grid Search.

Training Highlights

  • Dataset split: 80% Training, 20% Testing
  • 5-Fold Cross-Validation for robustness
  • Feature scaling using StandardScaler
  • Perfectly Balanced Dataset

Data & Features

The model is trained on a comprehensive dataset of over 70,000 patient records, considering critical health indicators.

Age
Gender
Height
Weight
Systolic BP (ap_hi)
Diastolic BP (ap_lo)
Cholesterol
Glucose
Smoking Status
Alcohol Intake
Physical Activity

Model Performance Visualization

Confusion Matrix Visualization

Detailed Analysis

The confusion matrix demonstrates low false negative rates, which is crucial for medical diagnostic tools to ensure no potential cases are missed.

Area Under Curve (AUC)0.80