Dosage planning for type 2 diabetes patients by ANFIS

Abstract

Diabetes Mellitus (DM) is a chronical disease that is increasing rapidly. Treatment of DM patients should be patient and careful. For DM patients, continuous medical supervision and dosage regulation is required to help them for living as healthy people. Usually the patients’ age, body mass index, genetic condition, blood and urine test results are used to decide to the appropriate drug doses by doctors. The aim of this study is management of drug dosages for DM patients by using data mining techniques. In this study, ANFIS and Rough Set (RSES) methods are used and compared for obtaining best benchmarking conditions. Input parameters consisted of gender, age, body mass index, genetics, insulin, fast blood glucose, urine, creatinine, cholesterol, triglyceride, HDL cholesterol, LDL cholesterol, uric acid, glucose in urine, acetone, diabetes type and dose of drugs. Dosage planning has been done for Glifix, Diamicron, Glucophage, Humilin_M, Novorapid and Insulin Lantus. As a result of this study, ANFIS give better results than RSES. ANFIS is the most effective data mining technique for dosage planning for Type 2 DM patients. © 2019 Elsevier B.V., All rights reserved.