Type 2 Diabetes Mellitus Diagnosis Model Using the C4.5 Algorithm

Abstract

Type 2 Diabetes Mellitus (DM) is a metabolic disorder characterized by elevated blood sugar resulting from decreased insulin secretion by pancreatic beta cells and/or impaired insulin function (insulin resistance). Over the last 50 years, there has been a rapid increase in the prevalence of diabetes, paralleling the rise in obesity rates. This study aims to develop a diagnostic model for type 2 DM using C4.5, incorporating feature selection and analyzing age and gender parameters of Type II DM patients. The research employs the Cross-Industry Standard Process for Data Mining (CRISP-DM). Based on the dataset used, the C4.5 model demonstrated superior performance compared to SVM and Random Forest, achieving an AUC value of 72.5%, indicating a reasonably good classification level. The predominant gender among Type II DM patients is female, comprising 210 patients or 54.8% in the age range of 18-94 years, while 173 male patients or 45.2% fall within the age range of 23-80 years.