DEVELOPING AN EXPERT SYSTEM FOR DIABETES MELLITUS PATIENTS

Abstract

Expert Systems are systems built to imitate human experts in problem solving and decision making. Medical Expert Systems are a breakthrough in medicine and proved beneficial in multiple critical areas such as Intensive Care Units (ICUs). Diabetes Mellitus is a chronic disease that affects a large population in the Middle East, and especially Saudi Arabia. To our knowledge, there is no medical expert system in the area of Diabetes Mellitus for the public and which supports the Arabic language. With this in mind, we aim to build a Diabetes Mellitus Expert System that is designed for the ordinary citizen to use. This expert system will be able to:  Perform correct screening,  Deduct if the patient is diabetic or not, if so, determines the type of diabetes, and its severity, or its complications and risks if not diabetic, and  Suggest appropriate treatment. Thus, the system will be a viable Arabic diabetes screening system that will also serve as a health education and awareness system. We are aiming to build this system using a hybrid rule-based approach that incorporates the use of Fuzzy Logic and frames for a powerful representations and accurate reasoning. A significant obstacle is the knowledge acquisition phase. Therefore several protocol generation and analysis techniques are employed to elicit the information from the medical domain which includes books and interviews. The results are verified by Endocrinologists who will also verify the accuracy of the system.