Advancing Cardiovascular Risk Prediction - A Review of Machine Learning Models and Their Clinical Potential

Aug 3, 2026

Upload by Quinn

This study conducts a systematic literature review of machine learning for predicting heart disease risk. Twenty recent articles are analyzed to identify commonly used algorithms and assess their performance, stability, and consistency across clinical datasets. Random Forest, Logistic Regression, Support Vector Machine, Decision Tree, and K-Nearest Neighbors are the most frequently applied models, with reported average accuracies of 89.56%, 83.14%, 83.14%, 82.57%, and 79.40%. The review discusses strengths, weaknesses, and challenges for clinical deployment to support medical decision-making.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages9

Reading Time23 min

AccessFree to download

Introduction

Research Methods

Article Selection

Results and Model Comparison

Advancing Cardiovascular Risk Prediction: A Review of Machine Learning Models and Their Clinical Potential Rona Regen*, Hendra Setiawan Master Program of Electrical Engineering, Faculty of Industrial Technology, Universitas Islam Indonesia Yogyakarta, Indonesia *Corresponding author, e-mail: [23925006@students.uii.ac.id](23925006@students.uii.ac.id) Abstract – This study conducts a systematic literature review on the application of machine learning technology in predicting heart disease risk. A total of 20 recent articles were identified and analyzed to evaluate the most used algorithms and their performance. The results show that Random Forest, Logistic Regression, Support Vector Machine, Decision Tree, and K-Nearest Neighbors are the most frequently applied models, with average accuracies of 89.56%, 83.14%, 83.14%, 82.57%, and 79.40%, respectively. In addition to comparing accuracy, this review also evaluates the strengths, weaknesses, and potential challenges of implementing each algorithm in clinical applications. The analysis reveals that RF demonstrates high stability and accuracy, making it the leading candidate for large-scale clinical heart disease risk prediction applications. These findings are expected to provide new insights for the development of more accurate, reliable, and clinically deployable machine learning predictive models to support medical decision-making. Keywords: Machine Learning, Heart Disease Risk Prediction, Clinical Applications, Predictive Modeling, Early Detection I. Introduction Cardiovascular diseases, including heart disease, are the leading cause of death worldwide, accounting for approximately 17.9 million deaths annually, or 32% of global deaths [1] . In this context, early detection and prevention of heart disease play a crucial role in reducing mortality rates. Machine learning (ML) technology has become a key tool in analyzing heart disease risks due to its ability to handle complex and large datasets, as well as its capability to produce more accurate predictions compared to traditional methods [2], [3] . This study aims to provide a comprehensive overview of the most frequently used ML models for predicting heart disease risk. We evaluate the performance of algorithms such as Random Forest (RF), Logistic Regression (LR), Supervised vector machine (SVM), Decision Tree (DT), and K-Nearest Neighbors (KNN), and analyze their strengths and limitations to provide guidance for clinical applications. While numerous studies have explored the performance of machine learning algorithms in predicting heart disease risk, this study offers unique contributions in several key aspects. It not only compares the accuracy of widely used algorithms such as RF, LR, and SVM, but also evaluates their stability and consistency across diverse clinical datasets. Additionally, this study emphasizes the strengths and limitations of these models, focusing on their suitability for real-world clinical applications. Clear visual comparisons are also provided to help readers better understand the findings and their practical implications. II. Research Methods This study employs a structured methodology focusing on the selection of relevant articles and the analysis of their data to examine the application of machine learning in predicting heart disease risk. II.1. Article Selection To identify relevant studies, a comprehensive search was conducted using Google Scholar with the keywords “Predictive Modeling Heart Disease Machine Learning.” The initial search yielded a substantial number of articles. These were systematically screened to ensure only the most relevant and high-quality studies were included. The selection process followed a structured approach with several inclusion criteria. First, only fully accessible articles with complete datasets were considered. Second, studies had to be published in reputable journals or conference proceedings to maintain research quality. Third,

FAQs

What is the main goal of this study?

To provide a comprehensive overview of frequently used machine learning models for predicting heart disease risk, along with their performance and suitability for clinical applications.

Which machine learning algorithms are most frequently applied in the reviewed studies?

Random Forest, Logistic Regression, Support Vector Machine, Decision Tree, and K-Nearest Neighbors are reported as the most frequently used models in the selected articles.

How does the review evaluate algorithms beyond accuracy?

It also examines strengths, weaknesses, stability, and consistency across diverse clinical datasets, and highlights practical challenges for clinical implementation.

Aug 3, 2026

Upload by Quinn