Heart Disease Prediction and Risk Analysis by Machine Learning Techniques.

Cardiovascular disease remains one of the leading causes of global mortality, underscoring the urgent need for robust early detection systems. This study introduces an innovative machine learning framework for heart disease prediction using the Cleveland dataset (sourced from the UCI repository via Kaggle). We employ a Decision Tree classifier for its transparency and visual interpretability, and benchmark its performance against a Gaussian Naïve Bayes model. Our data pipeline includes rigorous preprocessing steps—median imputation, Min-Max normalization, and stratified train-test partitioning—to ensure high-quality inputs. Visual outputs, including a detailed Decision Tree diagram (Figure 1) and a Principal Component Analysis (PCA) projection (Figure 2), elucidate the underlying data structure and decision boundaries. Although the Gaussian Naïve Bayes model demonstrates higher accuracy, the Decision Tree’s clear decision paths offer invaluable insights for clinical applications. This work balances predictive performance with interpretability, paving the way for future research with hybrid models and expanded clinical feature sets.