Rough-Fuzzy Classifier: A System to Predict the Heart Disease by Blending Two Different Set Theories

Abstract

Heart disease prediction has been a significant research over the past decade, because the major cause of death worldwide is due to heart disease. To predict the heart disease, several researchers combined fuzzy technique with some other technique for efficient classification purpose, since the fuzzy is efficient only if proper fuzzy rules are given in the rule base. Here, we have introduced a rough-fuzzy classifier that combined rough set theory with the fuzzy set. The overall process of the rough-fuzzy classifier is divided into two major steps, such as (1) rule generation using rough set theory, and (2) prediction using fuzzy classifier. At first, reduct and core analysis is used to identify the relevant attributes and the fuzzy rules are generated from the rough set theory after forming the indiscernibility matrix. Then, the fuzzy system is designed with the help of fuzzy rules and membership functions so that the prediction can be carried out within the fuzzy system designed. Finally, the experimentation is carried out using the Cleveland, Hungarian and Switzerland datasets. From the results, we ensure that the proposed rough-fuzzy classifier outperformed the previous approach by achieving the accuracy of 80 % in Switzerland and 42 % in Hungarian datasets.