Performance Evaluation and Comparison of Machine Learning Algorithms in Classification of CDC Diabetes Health Indicators Dataset by WEKA

Diabetes Mellitus is a chronic disease that occurs when the body loses its ability to produce insulin. When left uncontrolled and untreated, the sugar that accumulates in the veins causes complications that will reduce the quality of life and shorten life expectancy. Early diagnosis is important in diabetes to prevent these complications. In this study, indicators that play an important role in the early diagnosis of diabetes were classified into 3 different groups such as medical indicators, behavioural/lifestyle and overall indicators which combine two groups and machine learning methods applied to dataset groups for successful classification of diabetes. As a result, J48 performed 84.58 % accuracy for the classification of the medical indicator group, while SVM performed 84.05 % accuracy for the behavioural/lifestyle indicator group and Random Tree performed 84.25 % accuracy for the overall dataset. The results obtained showed that behavioural/lifestyle data are equally important as medical data in the early diagnosis of diabetes.