A Diabetic Prediction System Based on Mean Shift Clustering

An abnormal rise in glucose levels may lead to diabetes. Around 30 million people are diagnosed with this disease our country. In perspective Indian Council of Medical Research funded by Registry People diabetes India have taken an initiative and come up numerous solutions but unfortunately neither them has shape. Initially, the behavior chemical reaction between agent is estimated tracked region interest via mean shift algorithm using spatial range information. This color change related plasma concentration (plas), diastolic blood pressure, (pres.) Triceps skin fold thickness(skin), 2_hour serum insulin(insu), Body mass index age. These features obtained from these 768 instances classified Naïve Bayes Algorithm. The results compared previous work, integrated system K means approach terms sensitivity, specificity, precision, F-measure. It worth noticing that integration mean-shift clustering classification gives promising utmost accuracy rate 99.42% even after removing nearby duplicates predefined clusters.