Performance Analysis of Diabetes Prediction by using different Machine Learning Algorithms

Nowadays, daily life people are living bad lifestyle more having habit to consume more fast food and lack of health awareness. Fast food usually contains more sugar, fats, and too much oil which is the main cause of the increase of diabetes patient nowadays. Diabetes contributes to heart disease, kidney damage, nerve damage, eye damage, hearing impairment. To detect diabetes patient need to test their various bloods sample and are require to visit their nearby diagnostic center to get their report after consultation. These tests are too expensive and also takes too much time during test while diagnosis of large number of peoples. Machine learning algorithms help to predict whether a person is diabetic or non-diabetic. This paper’s approaches to detect diabetes risk of patient using medical data with the help Logistic Regression, Decision Tree, Linear SVM, Random Forest, Gradient Boosting, XG boosting with better accuracy. We are using 20 % of dataset for testing purpose and 80 % for training purpose. The analysis result shows that gradient boosting classifier had achieved highest accuracy than other classifiers which 79 % accuracy