Comparative Analysis of Classifier Models for the Early Prediction of Type 2 Diabetes

Abstract

Diabetes is a common disease which causes a large number of deaths each year and a large
number of people living with the disease do not realize their health condition early enough. If diabetes
is undiagnosed at early stages, it may result in high blood sugar, which lays a path to complications
like diabetic retinopathy, nephropathy, neuropathy, cardiac stroke, and foot ulcer, etc. So, the early
detection of diabetes is very important for the improved treatment, life quality and life expectancy of
patients. The traditional way of testing and diagnosis by the physician is not adequate to detect
diabetes. Nowadays, machine learning techniques are widely used for the automatic analysis and
prediction of diseases at an early stage using high dimensional biomedical data. In this paper, the
Pima Indians Diabetes Database (PIDD) sourced from UCI machine learning store is considered for
the investigation. Further, the most famous machine learning algorithms, decision tree, logistic
regression, random forest and SVM algorithms are applied for the prediction of diabetes. The
performance metrics of these algorithms are presented and compared which may be useful to assist
physicians with treatment decisions.