Type-2 Fuzzy Rule-based Expert System for Ankylosing spondylitis Diagnosis

Abstract

Ankylosing spondylitis (AS) is a chronic inflammatory disease. The pathogenesis of AS is poorly understood. However its association with human leukycyte antigen (HLA)-B27 is known. AS causes significant pain, disability, and social burden around the world [1]. Early diagnosis and treatment of AS are necessary in order to prevent or reduce all types of costs associated with loss of function. If the diagnosis is missed, however, the quality of patient's life will degrade. Besides, as a positive family history of AS is a strong risk factor for the disease, this negligence will put other family member's in jeopardy. Nowadays, Expert systems play a big role in diagnosis of patients with different diseases. The application of expert system to diagnose diseases started in the 70s with the development of Mycin. Expert systems in medical diagnosis can help in storing more knowledge than before and make it accessible in absence of a specialist and increase distribution of expertise. Our goal in this paper is to design a type-2 fuzzy rule-based expert system for AS diagnosis where the rules are evidence-based. The basic aim of evidence-based practice is to establish a narrow set of criteria for diagnosis based on research studies. System has mainly two parts. Firstly, the suspicion of disease is assessed for the persons identity according to odds ratio studies and patient's family history. Then the modified New York criteria (1984) for exploring sign and symptoms and the HLA-B27 examination result are considered. The system benefits from fuzzy reasoning and can manage the uncertain inputs through fuzzifying them and making use of type-2 fuzzy rules. Moreover, the system is connected to a spreadsheet for storing the patient's input data and system's final diagnosis. This system can be used by a non-rheumatologist in the diagnosis of AS or by a rheumatologist as an assistant.