A Comparative Review of Prediction Methods for Pima Indians Diabetes Dataset

Abstract

Diabetes Mellitus is generally considered to be a health issue that affects people and results in different types of problems like heart problems, vision problems, leg amputation and failure of kidneys if the disease diagnosis is not carried out in the perfect time. As Type II Diabetes Mellitus is a kind of abnormal disorder epidemic. The method of mining the medical information is concentrated in the recent days of research in the extraction of the significant data. This data is useful for the experts in the medical field, the enhancement of treatment and diagnosis of diabetic disorders. In this study, the review of different classification techniques and its details are implied on the dataset of benchmark Pima Indian diabetes. Neural Network (NN), Support Vector Machine (SVM), Machine learning algorithms is the different techniques of prediction that are used to predict Diabetics. This review also explains various advantages and disadvantages of the prediction methods and correlates the accuracy level of data classification. This also analyses the survey of different issues of recognizing the importance of relationship among the significant factors that lead to the growth of diabetics. The results prove that the data which is pre-processed provides a better accuracy level in classification. To validate the effectiveness of different models of prediction, we have precisely implied to Pima Indians diabetes datasets.