A Rule Based Expert System for Autism Diagnosis/Screening: Prototype Development
Author(s): Amina Sani Adamu, Abdullahi Saleh El-Yakub, Moussa Mahamat Boukar and Senol Dane*
Abstract
Autism is one of the challenging neurodevelopment disorders affecting individuals from childhood up to their adulthood. Most individuals with ASD have issues with social communication, and social interaction, exhibit some repetitive behavior, and have sensory issues. According to the DSM5 criteria, there are 3 severity levels of autism: level 1, level 2, and level 3 depending on the individual’s behavioral features. Diagnosing and screening autism is very challenging in terms of cost, the time it takes and the lack of enough experts to do it, especially in the low medium-income countries (LMIC) such as Nigeria. In this study, we have developed a prototype of a mobile-based expert system named Autism Tracker which can be used for parents and caretakers for screening autism. Primary healthcare workers and special needs educators can also use it for Diagnosis. The system can also be used for knowing the severity level of autism in the individual after diagnosis. It is a rule-based system where JRip classification algorithm was used for pruning rules used in developing the expert system for both diagnosis and severity. This expert system is in English language and has also been translated into three of the major languages spoken in Nigeria and some West African countries.