Early and Accurate Prediction of Heart Disease Using Machine Learning: a feature selection رسالة ماجستير

Abstract

The goal of this study is to determine the early and accurate prediction of heart disease using machine-learning techniques and feature selection methods. The study develops a machine learning-based technique for predicting heart disease. The study uses data gathered from Al Razi hospital in city of Jenin was 890 records, and because the sample size was insufficient to get accurate prediction, the sample size was increased by using the Generative Adversarial Network (GAN) Algorithm. The GAN algorithm generates synthetic samples that are added to the original dataset to enhance the size of the dataset and boost sample diversity. Simultaneously, feature selection approaches were used to determine the most significant features for heart disease prediction. To accomplish the study's results, machine learning methods, notably K-Nearest Neighbors (KNN), Random Forest, Adaboost, and Support Vector Machine (SVM) were applied to the selected features to create predictions. The results of this study show that the identified approach can achieve 99% accuracy by employing KNN and SVM models, and that using GAN-generated samples and feature selection approaches can improve the performance of machine learning models. These findings show that the approach that, combines feature selection and machine learning algorithms is useful for the early and accurate prediction of heart disease.