Heart Disease Risk PredictionStatistical Analysis and Classification of Heart Diseases Risk Using Clinical Parameter

Abstract

Heart disease remains a leading cause of morbidity and mortality worldwide, necessitating robust statistical methods for early detection and risk prediction. This study applies multiple statistical and machine learning techniques, including Logistic Regression, Random Forest, and Neural Networks, to analyze a clinical dataset of 1,025 observations with 14 variables related to demographic, physiological, and biochemical parameters. The study evaluates model performance using accuracy, sensitivity, specificity, and AUC metrics, while also assessing multicollinearity, correlation structures, and variable significance through standardized coefficients and information criteria (AIC/BIC). The Logistic Regression model achieved an AUC of 0.83, indicating strong predictive capability, whereas ensemble methods required parameter tuning to improve specificity. The results highlight key risk factors including chest pain type, ST depression, maximum heart rate, and number of major vessels, offering data-driven insights for preventive healthcare strategies.