Classification of Diabetes Mellitus Based on Machine Learning Classifier Techniques

In recent days diabetes is recorded as fastest growingst common disease among several diseases in the world. It became one of the major health problems in several states and countries. This occurs mainly when the normal human body is incapable to produce the sufficient amount of insulin in order to adjust the quantity of sugar levels in the body. This improper maintenance of sugar levels may lead to other diseases like heart disease, kidney disease, blindness, nerve damage and blood vessels damage. every one knows that there are mainly two general reasons for diabetes: One reason is the pancreatic gland not able make sufficient insulin or the body not produce make enough insulin. This type of symptom is mainly found on 5-10 % of citizens with diabetes and they come under Type-1 Diabetes. Another reason is cells do not respond to the insulin that is produced and this type of symptom people come under Type-2 Diabetes. In recent days, machine learning as well as DM techniques have been considered to design automatic diagnosis system for diabetes. In this proposed paper we aim to use the SVM, a ML method as a classifier for identification of diabetes data set. Here we applied the data cleaning techniques to handle the incomplete by fill in absent values, smoothening the noisy data, identified and removed the inconsistencies. By performing the data cleaning activities for verifying all the fields in the data set are properly arranged or not. Once after the data cleaning is completed then we try to apply SVM and Naive Bayes for classifying the diabetes dataset and we try to compare the both classifiers and find which classification techniques are efficient by comparing the both classifiers based on the results we observed. Our experimental results clearly tell that SVM can be effectively used for identifying diabetes disease. In this proposed application we try to analyze the diabetes based on location wise i.e. diabetes is differently affected by various people in and around the various corners of the city. In this proposed application we take sample data set of Visakhapatnam city with four area’s data like NORTH, SOUTH, and WEST & EAST. So by applying the SVM, we will try to analyze the different reasons which differ for each and every diabetes patients based on location wise. some area people mostly suffer with food, pollution, eating habits, daily habits, sleeping habits and other reasons. So based on each and every individual region diabetes differ one with other person in and around the city. So we are going to classify a set of patients data based on region wise and classify the cause of affecting diabetes for them and try to provide a counter measure and pre-cautions for the patients.