Evaluating machine learning models for prediction of coronary artery disease

Abstract

Coronary artery disease (CAD) is a prevailing global health issue and a leading cause of death worldwide. Its accurate and timely diagnosis is crucial for effectively managing the disease and improving patient outcomes. In this study, we conducted a comparative analysis of machine learning (ML)-based approaches to detect and diagnose CAD. A dataset of 918 instances from the UCI ML repository, comprising 11 typical risk factors and CAD predictors, was used for this investigation. The study deployed ML models in Google Colaboratory and PyCaret, testing their efficacy in diagnosing CAD. Our study provides a detailed overview of these ML methodologies, their strengths, and limitations, underscoring the potential of these algorithms to revolutionize CAD diagnosis and treatment. The overall goal of the study is to create a model that can predict the presence or chance of presence of CAD based on different parameters of the patient’s history. Findings include the showcased logistic regression model, which was proven to be particularly effective, with an area under curve of 0.88, indicating a high ability to differentiate between patients with and without CAD, and a successful ability to identify clinically key features of CAD such as the presence of exertional angina and chest pain. This study emphasizes the importance of further research in this field to establish ML as a cornerstone of modern healthcare diagnostics.