The Significance of Fine Tuning Parameters in Supervised Machine Learning Techniques for Diabetic Disease Prediction

Abstract

In health care analysis, data mining plays a significant role in disease prediction. Presently, our society large amount of death rates are due to diabetic disease. The mortality rate of the patients in diabetic disease has been increased every year. Researchers are dealing with various machine learning approaches which help the health care professionals to diagnose the disease in primary stage. Various classification mechanisms exist in the literature to predict the disease in primary stage. In this paper, the supervised machine learning algorithms namely Decision Tree, K Nearest Neighbor and Support Vector Machine are used for the prediction of diabetic disease. The purpose of this study is to emphasize the importance of hyper-parameter tuning to improving the performance of classifiers. The results obtained were compared with normal classifiers DT, KNN and SVM before fine tuning the parameters. From the results it is found that hyper parameters tuning improves the performance of the model.  Finally the results are evaluated by using various validation metrics..