Abstract

Info

Title:: Classification of diabetes based on K-Nearest Neighbor and neural network

Author(s):: CHEN Zhencheng1; DU Ying2; ZOU Chunlin3; LIANG Yongbo1; WU Zhiqiang4; ZHU Jianming1; 1. School of Life and Environmental Sciences, Guilin University of Electronic Technology, Guilin 541004, China; 2. School of Electronic Engineering and Automation, Guilin University of Electronic Technology, Guilin 541004, China; 3. Transforming Medical Research Center, Guangxi Medical University, Nanning 530021, China; 4. Medical Devices Testing Center of Guangxi Zhuang Autonomous Region, Nanning 530021, China

Keywords:: ; Keywords: diabetes; glycosylated hemoglobin; fasting blood glucose; K-Nearest Neighbor; neural network; cynomolgus monkey

Abstract:: Abstract: In order to achieve the early screening for diabetes and improve the accuracy of classification of diabetes, on the basis of studying the risk factors of diabetes, glycosylated hemoglobin is added as one of the features in the early screening for diabetes. Herein cynomolgus monkeys that are most similar to humans were selected as the study subjects. Several features, such as age, blood pressure, abdominal circumference, body mass index, glycated hemoglobin and fasting blood glucose, are chosen as inputs, while normal, prediabetes and diabetes are output as categories. Both K-Nearest Neighbor (KNN) and neural network are used to classify diabetes. When glycosylated hemoglobin is added as one of the classification features, the accuracy of classification using KNN (K=3) and neural network is 81.8% and 92.6%, respectively, significantly higher than the accuracy which is obtained without considering the feature of glycosylated hemoglobin (68.1% and 89.7%). Therefore, both KNN and neural network can classify and identify the data of cynomolgus monkey and achieve an early screening for diabetes.