Feature Selection Technique for Autism Spectrum Disorder

Abstract

Autism Spectrum Disorder (ASD) is a developmental disorder that restricts the development of behaviors, communication, and learning skills. People with ASD have difficulties in communicating and engaging with other people. Presently, a questionnaire called Autism Spectrum Quotient 10 (AQ-10) has been used to diagnose ASD symptoms for individuals [1]. However, using AQ-10 for ASD diagnosis is based on the observation from an individual's behavior and evaluate the result. It cannot show the outstanding attribute or relationship between ASD symptoms in different age-groups. Thus, this research proposed the novel feature selection technique for identifying the important attributes in AQ-10 results from three different age-groups, which are children at age of 7-12, adolescents at age of 13-20, and adults at age over 20. The results will represent the outstanding features and relationships between ASD symptoms in different age-groups, and lead to the proper design of ASD treatment programs for individuals.