Optimizing Hyper Parameters for Improved Diabetes Prediction

1Researcher, Imarticus Learning, Pune, Maharashtra, India 2 Assistant Professor, Singhad Institute of Management and Computer Applications, Pune, Maharashtra, India ---------------------------------------------------------------------***--------------------------------------------------------------------Abstract Diabetes also referred to as diabetes mellitus is a long-term condition that causes high blood sugar levels in an individual. The elevated blood sugar levels can damage the body in ways unimaginable, starting from infection risk, skin complications, heart diseases, strokes, kidney damage, nerve damage, blindness etc. Hence, correct diagnosis of diabetes mellitus is an important classification problem, the results of which will have a tremendous impact in the medical field. The purpose of this study is to emphasize the importance of hyperparameter tuning in improving the accuracy of data mining results. The results obtained were compared with KNN and ANN techniques recently proposed by other researchers and we came to a conclusion that tuning hyper-parameters can increase the accuracy of the model.