A Computational-Intelligence Based Approach to Diagnosis of Diabetes Mellitus Disease
Diabetes Mellitus (DM) is a disease that occurs when the pancreas cannot produce enough insulin or when insulin that it produces cannot be used effectively. High frequency of urination and hunger and thirst are general symptoms of high levels of blood glucose. Global estimates of 2015 claims that 415 million people are living with diabetes and 90% of them belongs to Type 2 DM.
DM have equal rates for men and woman, and a rate of 8.3% in total adults. Diagnosis of the disease is not challenging however, it requires blood glucose measurements in different times. In emergency cases where the patient is un-conscious, the possibility to overlook the disease is high. In this study, fuzzy c-means clustering algorithm, in which each variable can belong to more than one class, is used to classify the two groups of patients with and without diabetes through other blood test data and demographic factors. In the first application with 100 patients of a hospital, the algorithm correctly classified 81% of patients.
KeywordsFuzzy c-means clustering Unsupervised learning Medical decision support system
Authors would like to thank to Assoc. Prof. Esin Tuncay MD and Özlem Yılmaz Ünlü MD for their contribution. This study is financially supported by Galatasaray University Research Fund, project 18.402.004.
- 1.WHO - World Health Organization: Global report on diabetes (2016)Google Scholar
- 2.Dutta, S., Manideep, B.C.S., Basha, S.M., Caytiles, R.D., Iyengar, N.: Classification of diabetic retinopathy images by using deep learning models. J. Grid Distr. Comput. 11(1), 89–106 (2018)Google Scholar