Data Mining Approach for Diagnosing Heart Diseases through Deep Neural Network

Abstract

Predicting and detecting cardiac illness using machine learning and data mining approaches is extremely clinically useful, and further progress will be quite difficult. Heart disease early-stage prediction can be improved by using digital patient data to provide analytical support for clinical decision-making in countries throughout the world where cardiovascular knowledge is lacking and the number of erroneous diagnoses is high. Many supervised machine-learning algorithms were utilised in the study to find classifiers with the highest accuracy in predicting heart disease. They were also discriminated based on their performance and accuracy. The goal of the study is to use AI (Artificial Intelligence) to diagnose cardiac disease in both normal and pathological settings. Various AI technologies are expected to be used, with DNN (Deep Neural Network) outperforming the others. This is predicted by the analysis an updated spider monkey optimization (USMO) technique has been proposed for the DNN as a means of determining optimal weights. The investigation's findings reveal a precision of 96.77 percent in the Cleveland database and a precision of 100 percent in the Hungarian database.