COMPUTATIONAL MODEL FOR DIABETES PREDICTION USING DECISION TREE

Abstract

Diabetes is a long-term metabolic disorder characterized by elevated blood sugar levels that, over time, can seriously harm the heart, blood vessels, eyes, kidneys, and nerves. Around 422 million individuals worldwide have diabetes, the majority of whom reside in low- and middle-income nations and the disease claims 1.5 million deaths annually. Type 2 diabetes is the most frequent and most common in adults. This occurs when the body either cannot produce enough insulin or develops a resistance to it. The goal of this analysis is to create a system that can predict an individual's diabetic risk level based on their lifestyle and family history with greater accuracy by merging the results of various machine learning. To select the best model for diabetes prediction we have applied certain classification algorithms namely, Logistic Regression (LR), K-Neighbors Classifier (KNN), Support Vector Machine (SVM), Decision Tree Classifier, and Random Forest Classifier. This paper contains all the methods and techniques used in determining the best accurate model and also contains various visualization tools utilized to understand the data in-depth. Various experiments lead to the conclusion that the Decision Tree has an accuracy of 96.36%.