A Belief Rule Based Expert System to Assess Autism under Uncertainty

Autism is one of the most common neurological disorders found in children, resulting disabilities, which continue until adulthood. The accurate assessment of autism is considered as a challenging clinical decision making problem because of the presence of various types of uncertainties that exist with its factors such as social interaction, communication and behavior. These factors cannot be measured with 100% certainty since they are associated with various types of uncertainty such as vagueness, imprecision, randomness, ignorance and incompleteness. Consequently, traditional autism diagnosis procedures such as DSM-IV Criteria, Childhood Autism Rating Scale (CARS), Autistic behavior Interview (ABI) and Childhood Autism checklist for Toddlers (CHAT), which is carried out by a physician, is unable to deliver accurate result. Therefore, this paper presents the design, development and application of an expert system to assess autism under uncertainty. The Belief Rule Based Inference Methodology using the Evidential Reasoning (RIMER) approach, employed to develop this expert system. The knowledge base of this system constructed by using the real patient data as well as by taking expert opinion. Practical case studies were used to validate the expert system. The results generated from the expert system have been compared with the expert opinion as well as with the fuzzy rule based system. It has been observed that expert system's generated results are more effective and reliable than that of fuzzy rule based system and expert opinion.