Important Feature Selection & Accuracy Comparisons of Different Machine Learning Models for Early Diabetes Detection

More than 400 million people in the world have diabetes. High-risk factors of diabetic individuals vary dramatically, and many patients suffer complications and avoidable harm. Improving the identification level of high-risk factors would help to reduce the rate of complications. To do this, it is essential to analyze a person's medical record, detailed health information that currently requires doctors and is manual, time-consuming, and subjective. In this work, we introduce an approach to automatically predict type 2 diabetes mellitus (T2DM) applying a neural network. The objective of this paper is to find which type of model that works best for predicting diabetes. We used the Pima Indian Diabetes data-set in this analysis. The analysis was carried out on this database using two methods. The first method includes Data Recovery followed by feature selection. We input these features to the MLP neural network classifier which achieved an accuracy of 85.15%. In our second approach, we applied noise reduction based method using k-means followed by feature selection. The features thus obtained are used with Random Forest, Logistic Regression and MLP neural network classifier. The maximum accuracy obtained among these classifiers is 77.08%. The consultation shows why Data recovery with MLP is far better than K-means based noise reduction with the different type of classifier.