A novel approach for diagnosing heart disease with hybrid classifier

Abstract

A Cardiovascular disease which is additionally referred to as heart diseases that have been a common and steady issue in the field of medical research. Nowadays various methods were applied which is not robust for the prediction of human being expenses and disease risks for patients. This paper proposed an Orthogonal Local Preserving Projection (OLPP) method to reduce the function dimension of the input high-dimensional data. The dimension reduction improves the prediction rate with the help of hybrid classifier i.e. Group Search Optimization Algorithm (GSO) combine with the Levenberg-Marquardt (LM) training algorithm in the neural network. The LM training algorithm is used to solve the optimization problem and it determines the best network parameters such as weights and bias that minimizes the error. The final output of the optimization technique is combined with the performance metrics as accuracy, sensitivity, and specificity. From the result, it is observed that hybrid optimization techniques increase the accuracy of the heart disease prediction system.