Diabetes Mellitus Prediction and Classification with a Randomizable filtered Classifier Ensemble Algorithm of the K-Nearest Neighbor

Abstract

Diabetes is a chronic and life-threatening condition caused by high levels of sugar glucose in the blood, which primarily affects the elderly in all societies. This has become a huge global health issue. Patients, diseases, causes, and medical services are all collected in vast amounts by medical care systems and hospitals. Diabetes has been the subject of numerous studies aimed at detecting and preventing the disease. This study presents an Ensemble of Randomizable Filtered Classifier K-Nearest Neighbor Algorithms as a data mining tool for predicting patients who are prone to develop diabetes mellitus. The Pima Indian Diabetes Dataset was used, which contains information on patients with diabetes and the possibility of acquiring diabetes. For design and implementation, the study used an ensemble of randomizable filtered classifiers K-Nearest Neighbor Algorithm and WEKA Software. The proposed approach predicts the likelihood of developing diabetes mellitus, according to the findings. It is suggested that the plan be implemented by the health-care system in order to discover diabetes mellitus patients early.