AI-DRIVEN DIAGNOSIS OF HEART CONDITIONS: A COMPARATIVE STUDY ACROSS NEURAL NETWORKS AND TRADITIONAL MODELS

Abstract

Timely prediction of cardiac illnesses is crucial for life preservation. Traditional approaches to prediction have often struggled with issues including increased prediction cost, longer calculation times, and increased complexity with increasing data volumes, all of which have a negative impact on forecast accuracy. For this reason, several ML (Machine Learning) and DL (Deep Learning) algorithms have been included into AI (Artificial Intelligence) systems for the purpose of cardiac illness diagnosis. Improved detection is a result of its ability to learn from large datasets that include patients' ages, weights, blood pressures, and other risk variables, and then extract relevant information based on user input. The analysis of illness incidence from historical data is substantially aided by the storing of greater data using AI. To that purpose, this study will survey the literature on artificial intelligence (AI) based algorithms used for cardiac illness prognostication and highlight its merits. It evaluates and compares things in the same way that conventional research in the field has, with an emphasis on accuracy and making the most of algorithms. Aims of future research include identifying the dimensions with high and low prediction accuracies so that relevant studies can be conducted, with the paper's main findings highlighting the development and ongoing exploration of AI techniques for heart disease prediction. The article concludes with a section on future research that should serve as a catalyst for more studies into the use of AI in the identification of heart diseases.