An Ensemble Approach of Supervised Learning Algorithms and Artificial Neural Network for Early Prediction of Diabetes

Diabetes mellitus, is a long-term illness that impairs the body's ability to absorb sugar or glucose. The presence of glucose in the circulation can increase if diabetes is not treated progressively and cautiously, providing a health risk for hypertension and arteriosclerosis. Type 1 and Type 2 diabetes are, in fact, the two major kinds. If an individual does not generate enough insulin to fulfill the body's needs, they will acquire type 1. Type 2 diabetes affects the way human body uses insulin (insulin resistance). Medical datasets may be used to apply machine learning to identify and disease prediction in a more robust and appropriate approach. This analysis is based on implementing machine learning algorithms on a publicly available two datasets that contains signs and symptoms suggesting if an individual is diabetic or not. The dataset is explored with the outcomes of machine learning models Logistic Regression, KNN, AdaBoost, and Multilayer Perceptron. Finally, StackingCVClassifier is used to ensemble these four models. And at last, the study revealed that alone Multilayer perceptron has the best accuracy, with 91 percent and 93 percent for two distinct datasets, respectively.