Enhanced Prediction of Heart Disease Using Particle Swarm Optimization and Rough Sets with Transductive Support Vector Machines Classifier

Abstract

Over the last decade heart disease has significantly increased and it has emerged to be the primary reason behind the mortality in people living in many nations across the world. The computer-assisted systems act as a tool for the doctors in the prediction and diagnosis of heart disease. In the medical domain, Data Mining yields a variety of techniques that are extensively employed in the medical and clinical decision support systems that has to be quite useful in diagnosing and predicting the heart diseases with less time and good accuracy to improve their health. The previous system designed a radial basis function with support vector machine for heart disease prediction. However it does not provides a satisfactory classification result. To solve this problem the proposed system designed a Particle Swarm Optimization and Rough Sets with Transductive Support Vector Machines (PSO and RS with TSVM) based prediction is performed. In this proposed work, the dataset of the heart disease is collected from UCI repository. In order to reduce data redundancy and improve data integrity, the data normalization is performed by using Zero-Score (Z-Score). Then Particle Swarm Optimization (PSO) algorithm and Rough Sets (RS) based attribute reduction technique is used for selecting the optimal subset of attributes that, in turn, minimizes the computational hurdles and improves the performance of the prediction system. Finally, the Radial Basis Function-Transductive Support Vector Machines (RBF-TSVM) classifier is used for heart disease prediction. The results obtained from the experiments indicate that the system proposed accomplishes a superior performance in comparison.