COMPARATIVE ANALYSIS OF MACHINE LEARNING MODELS FOR PREDICTING HEART DISEASE

Abstract

Cardiovascular diseases, particularly heart disease, remain among the leading causes of morbidity and mortality globally [1]. As the prevalence of heart-related conditions increases, there is a growing demand for early diagnostic systems that can support clinical decision-making and preventive care [2]. This whitepaper presents a comparative analysis of multiple machine learning algorithms applied to the UCI Heart Disease dataset [3], leveraging both statistical and predictive modeling approaches to identify patterns associated with heart disease. The study begins with an in-depth exploratory data analysis (EDA) to uncover trends, outliers, and correlations among clinical attributes such as age, cholesterol levels, resting blood pressure, and electrocardiographic results [4]. Following EDA, a suite of machine learning models—including Logistic Regression, Random Forest, and Gradient Boosting—are implemented to classify patients based on the likelihood of heart disease presence. Each model is evaluated using robust metrics including accuracy, precision, recall, F1-score, and the area under the ROC curve (AUC), enabling a performance-driven comparison [5]. Our findings indicate that ensemble-based models such as Gradient Boosting and Random Forest consistently outperform baseline models in predictive accuracy and sensitivity, making them ideal candidates for integration into clinical diagnostic tools [6]. The insights from this study highlight the critical role of feature selection, preprocessing, and model interpretability in healthcare AI applications [7]. This work contributes to the ongoing advancement of data-driven health technologies by demonstrating the potential of machine learning in enhancing the early detection and risk stratification of heart disease.

In the broader context, the United States' healthcare expenditure underscores the urgency for efficient diagnostic tools. In 2023, U.S. healthcare spending reached $4.9 trillion, accounting for 17.6% of the nation's Gross Domestic Product (GDP), with per capita spending at $14,570 [8]. Notably, hospital care, physician and clinical services, and retail prescription drugs collectively accounted for 60% of total spending [9]. Despite this substantial investment, the U.S. continues to face challenges in achieving optimal health outcomes, particularly in managing chronic diseases like heart disease. Implementing effective machine learning models for early detection can play a pivotal role in improving patient outcomes and reducing healthcare costs [10].