Diabetes screening using machine learning algorithm.

Currently, health management is a major concern all over the world. Particularly in India, due to population, the health issues are increasing rapidly. Diabetes has become a major issue that affects people all over India. With the advancement of Machine Learning approaches, many solutions have been developed to curb this serious issue. The goal of this paper is to design a model that can detect diabetes at preliminary stages with more accuracy. Therefore, various classification algorithms namely K-Nearest Neighbor (KNN), Support Vector Machine (SVM), Logistic Regression (LR), Decision Tree (DT), Random Forest (RF), and Gradient Boosting (GB) are used to solve this problem at earliest. Various performance metrics like accuracy, sensitivity, specificity, precision and recall are used for evaluation of the model. Evaluations are performed on PIMA Indian diabetes database, which is pulled from the Machine Learning cache. Results procured manifest that Random Forest outplay with the highest precision of 85.06%. The results are appropriately authenticated through a bar graph representation. [ABSTRACT FROM AUTHOR]