A COMPARATIVE ANALYSIS OF SOME SELECTED MACHINE LEARNINGALGORITHMS FOR CLASSIFICATION AND PREDICTION OF DIABETES TYPES

Abstract

Diabetes mellitus is one of the most common human diseases worldwide and may cause several health-related complications. It is responsible for considerable morbidity, mortality and economic loss. A timely diagnosis and prediction of this disease could provide patients with an opportunity to take the appropriate preventive and treatment strategy. To gain an understanding of risk factors, the study explores the comparison of six classification algorithms along with the artificial neural network algorithm in the prediction of diabetes types. The analysis of data from patient records shows four main predictors of type 1 & type 2 diabetes: Age, Ulcer Duration, White Blood Cell (WBC) and Erythrocyte Sedimentation Rate (ESR). All the classification algorithms based on accuracy, precision, recall, and F-measure show good results for the parameters with accuracy greater than 95%. However, random forest and k-nearest neighbor algorithms provided 99% accuracy for the train/test split. Thus, these models can be applied to make a reasonable classification of type 1 and type 2 diabetes. Hence, the study ensures a timely diagnosis and prediction of this disease for appropriate preventive and treatment strategy.