Utilizing Machine Learning for the Early Detection of Coronary Heart Disease

Abstract

Coronary Heart Disease (CHD) is a persistent health issue, and risk prognosis is very important because it creates opportunities for doctors to provide early solutions. Despite such promising results, this type of analysis runs into several problems, such as accurately handling high-dimensional data because of the abundance of extracted information that hampers the prediction process. This paper presents a new approach that integrates Principal Component Analysis (PCA) and feature selection techniques to improve the prediction performance of CHD models, especially in light of dimensionality consideration. Feature selection is identified as one of the contributors to enhance model performance. Reducing the input space and identifying important attributes related to heart disease offers a refined approach to CHD prediction. Then four classifiers were used, namely PCA, Random Forest (RF), Decision Trees (DT), and AdaBoost, and an accuracy of approximately 96% was achieved, which is quite satisfactory. The experimentations demonstrated the effectiveness of this approach, as the proposed model was more effective than the other traditional models including the RF and LR in aspects of precision, recall, and AUC values. This study proposes an approach to reduce data dimensionality and select important features, leading to improved CHD prediction and patient outcomes.