Diabetes Disease Prediction Using Machine Learning Algorithms
Arwatki Chen Lyngdoh, Nurul Amin Choudhury, Soumen Moulik
Abstract
This paper deals with the prediction of Diabetes Disease by performing an analysis of five supervised machine learning algorithms, i.e. K-Nearest Neighbors, Naïve Baye, Decision Tree Classifier, Random Forest and Support Vector Machine. Further, by incorporating all the present risk factors of the dataset, we have observed a stable accuracy after classifying and performing cross-validation. We managed to achieve a stable and highest accuracy of 76% with KNN classifier and remaining all other classifiers also give a stable accuracy of above 70%. We analyzed why specific Machine Learning classifiers do not yield stable and good accuracy by visualizing the training and testing accuracy and examining model overfitting and model underfitting. The main goal of this paper is to find the most optimal results in terms of accuracy and computational time for Diabetes disease prediction.