Early prediction of diabetic type 2 based on fuzzy technique

Abstract

Intelligent analysis of present lifestyle may help to understand the development of the chronic diseases and the relationship of these diseases together. It is possible to reduce or prevent the development of these diseases. In this work, a novel intelligent method is introduced and applied for early detection of type 2 diabetic. Intelligent analysis depends mainly on evaluation life-threatening conditions (obesity, hypertension, smoking status, alcohol drinking status and low level of physical activities) to extract knowledge from linguistic variablesand design a new cognitive tool to assist in the prediction process.This method consists from three stages: in the first stage, data was collected from 100 healthy volunteers, which includes evaluations of life-threatening conditions. The second stage is implementation of fuzzy model for early prediction of type 2 diabetes. Predicted blood glucose values of proposal technique were compared with average fasting blood glucose values based on analysis of Bland-Altman plot. Furthermore, fuzzy system model presents superior results (accuracy = 81%, precision = 0.57% and recall = 0.83%).