Healthcare Diagnosis Support System for Detection of Heart Disease in a Patient using Machine Leaming Methods

Abstract

One of the most considerable investigative areas has remained the applications area of medical advancement. The early warning method for heart disease (HD) is one of these medical technologies. The goal of a healthcare diagnosis support system (HDSS) is to diagnose HD at an early stage such that the diagnosis can be streamlined, advanced cases stopped, and care costs can be minimized. A machine learning (ML) HDSS for heart disease identification is obtainable in this study, and it is capable of obtaining and learning information from each patient's experimental data automatically. The authors employed a dimensionality reduction technique autoencoder (AE) with three ML classifiers detection of HD. The HD dataset employed for the HDSS was collected from the National Health Service (NHS) database. The result was evaluated using the confusion matrix performance measures such as accuracy, specificity, detection rate, Fl score, and precision. The result shows that NB+Autoencoder outperformed the other two classifiers with an accuracy of 57.2% and 55.4 precision.