Integration of Rough Set theory and Genetic Algorithm for Optimal Feature Subset Selection on Diabetic Diagnosis

Diabetic diagnosis is an important research in health care domain to analyze relevant microorganisms at an earlier stage. Due to large growth in world’s population, feature subset selection model receives a great deal in any domain of research and also a reliable tool for diabetic diagnosis. Several data mining techniques have been developed to evaluate the significant causes of diabetes with least sets of risk factors. These minimum set is selected without considering the potential significance of the risk factors and optimal feature subset selection, hence it failed to diagnosis the pattern of diabetes accurately. In order to improve the feature subset selection, an Integration of Fuzzy Rough Set Theory and Optimized Genetic algorithm (IFRST-OGA) is introduced. The main objective of the IFRST-OGA is to find optimal risk factors for efficient pattern recognition on diabetes healthcare data. Initially, feature selection is performed using Fuzzy Rough Set Theory (FRST) for diagnosing the diabetes. After that, the Optimized Genetic Algorithm (OGA) is applied which mainly searches for an optimal feature subset through the selection, crossover, and mutation operations to diagnose the disease at an earlier stage. This helps to identify the risk factor and diagnosing the diabetes disease efficiently. Experimental results show that the proposed IFRST-OGA increases the performance in terms of true positive rate, computation time and diabetes diagnosing accuracy.