An Optimized Random Forest Classifier for Diabetes Mellitus

Abstract

Machine learning-based classification algorithms help in diagnosing the symptoms at early stages by prior diagnosing of symptoms and taking medications according to it. Combining of genetic algorithm with the random forest classifier can optimize the results obtained only by the random forest classifier. In this proposed system, genetically optimized random forest classifier is used for the classification of diabetes mellitus. Aims. To develop an optimized random forest classifier by genetic algorithm for diabetes mellitus. Methods. A genetic algorithm is used in the first stage for optimizing random forest, and the optimized outputs are fed into the fine-grained random forest to diagnose the symptoms of diabetes mellitus. Results. In this analysis, the proposal of hybrid optimized random forest classifier (GA-ORF) with a genetic algorithm is made. In this evaluation, the various performance metrics of classifiers, GA-ORF has achieved accuracy higher than of the previously proposed classifiers for diabetes mellitus.