Expert System for GERD Diagnosis Based on Fuzzy Logic at dr. Rumbang Sembiring, Sp.PD’s Practice

Abstract

Gastroesophageal Reflux Disease (GERD) is a chronic digestive disorder characterized by chest burning (heartburn), epigastric pain, nausea, regurgitation of stomach contents, difficulty swallowing, sleep disturbances, and chronic cough. Conventional diagnostic methods like medical interviews and endoscopy have limitations in terms of time, cost, and access. This research aims to design and implement an Expert System for GERD Diagnosis based on Mamdani Fuzzy Logic with the Forward Chaining method to assist in rapid, accurate, and efficient initial diagnosis. Primary data was collected from the practice of dr. Rumbang Sembiring, Sp.PD (Internist), in the form of intensity values for seven main GERD symptoms. The system was developed as a web-based application using Python and PHP programming languages, involving the stages of fuzzification, inference based on IF-THEN rules using forward chaining, aggregation, and defuzzification using the centroid method to produce a category of GERD likelihood (low, medium, high). Evaluation was conducted by comparing the system's results with the doctor's diagnosis on 40 test data points. The test results showed an accuracy rate of 87.5%, indicating that the system can represent expert knowledge very well. This system is expected to be an initial diagnostic aid for medical personnel and patients and can be further developed by adding symptom variables and mobile application integration