Abstract: Over 30 million people in India are affected by diabetes and lots of people are under the danger position. Thus, early diagnosis and treatment is required to avoid/prevent diabetes and its associated health problems. The medical data mining methods and techniques explored in this work help to identify the suitable techniques for efficient classification of diabetes datasets and to provide effective recommendations.The standard dataset obtained from Pima diabetes database is used for detecting proposed system. The data set contains data for 769 patients contains both sick and healthy patient’s data are obtained. The research work also performs the analysis of the features in the dataset and selects the optimal features based on the correlation values. The SVM algorithm and Random forest giving the highest specificity of 91.55% and 92.8%, respectively holds best for the analysis of diabetic data.

Keywords: Data mining, diabetics, KNN, decision tree, Jupiter note book, python 3, membership function

| DOI: 10.17148/IARJSET.2021.8208