Prediction of Autism Spectrum Disorder in Children: A Detailed Systematic Review on Machine Learning and Deep Learning Methods

Abstract

Autism spectrum disorder (ASD) is a serious, chronic neurodevelopmental illness characterized by developmental difficulties that are permanent or restrict the growth of thinking, behaviour, activities, and social-communication skills. The signs of autism are more pronounced and more straightforward to identify in youngsters between the ages of two and three. Although there is no permanent cure for ASD, it is still challenging to identify meltdowns or other difficulties in the early stages of medical care. In order to promote brain development and raise awareness of ASD among parents and caregivers, the survey’s objective was to identify ASD at an early stage. Methods like deep learning (DL) and machine learning (ML) are currently employed to predict autism spectrum illnesses. This study provides an in-depth review of papers that predict ASD using ML and DL, as well as data analysis and classification techniques. Additionally, this survey intends to categorize and examine the various ML and DL approaches, as well as to describe the characteristics of ASD, assess performance, and show the scope of future research. For upcoming academics who want to study ASD prediction using ML and DL, this publication offers a road map.