Study of Machine Learning Models to Detect Autism Spectrum Disorder

Abstract

ASD is a spectrum disorder, some people with ASD may have mild symptoms and can live independently, while others may need more support and assistance in their daily lives. Some of the risk factors for ASD include having an older parent, being born prematurely, having a sibling with ASD, or being exposed to certain infections or toxins during pregnancy some of the common types of treatments include behavioural therapy, occupational therapy, physical therapy, and passionate about their interests. Such as the Modified Checklist for Autism in Toddlers, observational tools, such as the Autism Diagnostic Observation Schedule, and developmental assessments, such as the Mullen Scales of Early Learning. This study refers to some Machine Learning (ML) based applications that are able to detect Autism among individuals. This comprehensive review explores the innovative integration of machine learning (ML) models in the detection and diagnosis of Autism Spectrum Disorder (ASD). It begins by highlighting the complexities and diagnostic challenges of ASD, noting the limitations of traditional assessment methods. The review then delves into the realm of artificial intelligence (AI), discussing how AI, particularly ML and deep learning techniques, are revolutionizing the approach to ASD detection. It covers various ML models, including supervised, unsupervised, and reinforcement learning, and their application using behavioural, genetic, and neuroimaging data. A significant focus is given to the use of Logistic Regression and Hybrid Autism Screening Models in predicting ASD. The review also examines the efficacy and performance of supervised and deep learning models in ASD detection, evaluating their accuracy and precision. By providing a detailed analysis of the current state of AI in healthcare, specifically for ASD, this review underscores the potential of ML models in offering more accurate, accessible, and efficient diagnosis methods for ASD