Deep Learning Model for Diagnosing Diabetes

Diabetes is one of the most prevalent chronic diseases among all the age groups in developing nations like India. The work collates various machine learning techniques such as SVM, random forests, naive bayes and proposed MLP to perform better detection of diabetes in humans. The goal of this study is to predict the diabetes through neural network of multilayer perceptron. In this study, multilayer perceptron of deep learning algorithm is compared with different machine learning algorithms. The model for the different algorithm is validated with fivefold cross validation. The machine learning algorithms in this study are support vector machine, random forest, and naive bayes. The dataset used for this study is taken from UCI machine learning repository. This algorithm is trained and tested with all the features of the dataset. The algorithms are evaluated based on the accuracy measures. The result obtained shows that the multilayer perceptron of deep learning algorithm gives an accuracy of 98%.