An Enhanced Diabetes Mellitus Prediction Using Feature Selection-Based Type-2 Fuzzy Model

The diabetesmellitus has been known to be a serious illness and revered for its ability to cause high mortality rate. This disease is famous among both youth and adult for its existence in the human body, and very difficult to diagnose, thus, produces an under-diagnosis issue when clinicians try to pinpoint the precise symptoms for disease prediction. The majority of the currently used diagnostic and monitoring methods are designed on type 1 fuzzy logic or ontology, which, as a result of the inconsistent and ambiguous nature of the collected data, is unsatisfactory. Therefore, this paper proposes an enhanced feature selection-based enabled type-2 fuzzy logic (T2FL) model for the prediction of diabetes patients. The proposed model used Particle Swarm Optimization to select the most relevant features from the dataset so as to remove irrelevant features from the data, and T2FL technique was used for the classification of the disease. The model extract precise information and correctly conclude the result. The proposed technique utilizes T2FL to determine the membership values of the clinical information, and the decision-making mechanism properly processes the evidence derived from the crisp values. A comprehensive computer simulation using a diabetes dataset shows that the suggested strategy works noticeably better than the ones already in place based on type-1 fuzzy logic and ontology in terms of determination effectiveness.