Early Prognosis of Diabetes Using Supervised Learning Techniques: A Comparison of Performance

Abstract

Supervised machine learning algorithms have been a predominant technique in information mining field. Disease forecast utilizing health information has shown a potential application region for these techniques. This investigation centers to distinguish the vital patterns among various kinds of supervised learning and their performance and utilization for diabetes prognosis. The point of this work is to analyze the performance of various machine learning (ML) classifiers. These ML classifiers are k-Nearest Neighbors, Support Vector Machine, ANN, Logistic Regression, Decision Tree and Ensemble classifiers which are applied on diabetes datasets for evaluating performance. The ML models are trained and tested with Pima Indian Diabetes dataset. The exploratory outcomes uncover that the SVM classifier beat the different classifiers with a most noteworthy accuracy of 83% in detecting diabetes.