Investigation and Analysis of Autism Spectrum Disorder using Artificial Intelligence Techniques

Abstract

Neurological disorders generally affect the nerves of the brain and spinal cord in the human body. Autism Spectrum Disorder (ASD) is one of such neurological disorders which at present has become one of the most severe developmental disabilities that cause newlinesocial, communication, and behavioural changes in individuals. Early signs of ASD are observed during the first 6 to 18 months of an individual s life span. In due course of time, the symptoms are marked by developmental regression such as loss in social, verbal as well as communication ability within between 18 to 36 months of the individual s lifespan. Overall, ASD is characterized by repeated activities in addition to aimless imaginary thoughts. Early recognition of the disorder is one of the solutions to the problem so that precautionary measures can be adopted against the disorder. newlineTo get rid of the expensive diagnosis as well as time-consuming procedure for ASD, a mobile-based screening tool termed as ASDTest app was developed. The app collected over 1400 instances covering toddler, child, adolescent, and adult categories of individuals. It is publicly available on research sites like Kaggle as well as UCI Machine Learning repository. newlineThe Proposed research work in this thesis is based on the detection of ASD class types in newlineall categories of individuals. The main objective of the research work is to understand the characteristics and situations faced by individuals suffering from ASD and early diagnosis of it leading to improved developmental and adaptive functioning with enhanced social skills. Out of all categories, the toddler data set is found to be an unbalanced data set. newlineTherefore, in this thesis along with all categories, much emphasis is also given to the toddler data set. The original data sets are first pre-processed where they underwent newlinestandardization as well as dimension reduction.