A Fuzzy Logic Approach for Predicting Gestational Diabetes Mellitus Using Risk Factors
Nowshad Hasan, Md. Saiful Islam, Md. Tanvir Hayat
Abstract
Gestational Diabetes Mellitus (GDM) is a health condition distinguished by heightened glucose levels experienced during pregnancy. Early detection of GDM is key to managing the condition and protecting both the mother and fetus. The aim of this study is to introduce a fuzzy expert system that is designed to assist both medical professionals and non-specialists in the evaluation of patients with GDM using risk factors. The approach employed five risk indicators such as age, body mass index, family diabetes history, previous GDM history and poor obstetric outcome as inputs. Triangular and trapezoidal fuzzy numbers are applied in this process. The Mamdani Fuzzy Inference System is used, with specific rules established for GDM prediction. The results indicated three levels of gestational diabetes: low, medium, and high. The resultant fuzzy expert system demonstrated promising potential for supporting medical professionals during the decision-making process pertaining to GDM. In conclusion, while the developed fuzzy expert system for GDM prediction cannot replace the invaluable work of specialist doctors, it can serve as a supportive tool in healthcare settings.