A COMPREHENSIVE SURVEY ON HEART DISEASE PREDICTION USING MACHINE LEARNING AND DEEP LEARNING APPROACHES

Abstract

Globally, the death rate is increased by one of the major conditions named heart disease (HD). This HD greatly impacts the global healthcare systems. For the purpose of enhancing outcomes of the patient and reducing medical challenges, the early detection (ED) and diagnosis of cardiovascular disease (CVD) is crucial. Then, the implementation of the artificial intelligence (AI), namely machine learning (ML) and deep learning (DL) techniques have revolutionized the predictive modelling of HD. A comprehensive insights regarding the recent developments in HD prediction with the application of the machine learning (ML) and deep learning (DL) algorithms, including logistic regression (LR), decision trees (DT), support vector machines (SVM), random forests (RF), neural networks (NN), convolutional NN (CNN), and long short-term memory (LSTM) models was offered in this study. Here, the commonly utilized datasets, feature selection (FS) strategies, data pre-processing approaches are all examined in this study. Then the study also analyses the assessment metrics that will helps in determining the accuracy (ACC) and dependability of the predictive models (PM). The benefits, drawbacks, and efficacy of every model is identified by the comparison of models. This survey also facilitates in resolving issues like data imbalance, model interpretability, privacy issues, and practical deployment limitations. Recommendations regarding future directions, like explainable AI, federated learning (FL), and the integration of multi-modal (MM) health data was also offered in this study, and it may help the experts in creating more clinically valuable and dependable prognostic tools. This comprehensive survey contributes the scholars and professionals in creating intelligent systems for HD diagnosis and risk assessment. So, this comprehensive survey is beneficial.