Diabetes Prediction Using Machine Learning Techniques: A Comparative Analysis

摘要

Nowadays, Machine Learning and Artificial Intelligence play a important role in the healthcare sector. Diabetes is one of the most populated diseases in the world according to WHO. It is caused due to the increased level of glucose in the body. There are some more attributes on which diabetes can be predicted. This work mainly focuses on building diabetes aided system which can predict the disease at the earliest possible stage. In this paper, w e ised different ML techniques to predict diabetes at initial phases. In Machine Learning, support vector machine, logistic regression, Decision Tree, Random Forest, gradient boost, K-nearest neighbor, Naive Bayes algorithm are used. We measure these algorithms by using the following metrics (1) precision level, (2) accuracy level, (3) recall, (4) F-measure. The aim of this analysis is to compare different techniques to obtain better accuracy. It is observed that the Random Forest and naive base algorithm obtained an accuracy of 80%.