Prediction of diabetes disease using Adaboost, Random Forest Algorithm and Logistic Regression

Diabetes is sometimes called the silent killer as it takes long for the symptoms to appear. Diabetes is a common disease that can be prevented or postponed with early detection based on various features. In this paper, collect dataset from UCI repository having 17 attributes and 520 records of different patients. In addition to that, these attributes are pass to machine learning techniques like Random forest, Adaboost and Logistic regression, which is used to predict diabetes disease. The performance is measured in terms precision, recall and F-measures. It has been observed, during training phase random forest achieved 100% and testing phase 99.36% accuracy. The attained results seem to be more accurate and precise, as compared to existing techniques.