HEART ATTACK RISK PREDICTION WITH WEKA

Abstract

The thesis presents a solution to the classification problem to improve the performance of artificial neural networks and the development of functions for improving a well-known dataset for predicting heart attacks. The study aims to increase the predictive power and reduce the size of the dataset using deep learning methods for clinical diagnostic purposes. It is possible to improve the accuracy of heart attack risk prediction by reducing the size of the raw data set. This leads to more reliable, less time-consuming and less expensive clinical diagnoses. Optimal dimensionality and noise reduction in the heart disease dataset and the application of classification algorithms using artificial neural networks are promising to improve its predictive ability to a clinically acceptable standard.