Performance Analysis of Machine Learning Algorithms with Hyperparameter Tuning for Diabetes Prediction

Abstract

All over the world, many people are affected by Diabetes. At present in hospitals various blood and urine tests were carried out to find whether a person is having Diabetes or not. Nowadays Machine learning plays a major role in the biomedical areas in prediction and identification of various diseases. In this paper, a detailed performance analysis about the prediction of Diabetes using various machine learning algorithms is presented. Performance of Linear Regression, Decision Tree, XGBoost, Support Vector Machine and K Nearest Neighbour algorithms are compared. Grid search and Randomized search hyperparameter tuning are also used to tune the hyperparameters of all these machine learning models. The PIMA Indians Diabetes Dataset is used for this work. The prediction results of these models are compared using Accuracy, Precision, Recall and F1 score. The results of this work show KNN algorithm with randomized search hyperparameter tuning gives better results than all other machine learning models with grid search hyperparameter tuning.