Detection of Autism Spectrum Disorder in Toddlers using Machine Learning

Abstract

The aim of this study is to identify
toddlers at risk for Autism Spectrum Disorder (ASD)
early on by developing a web-based tool that uses the
machine learning method logistic regression. Our
approach emphasises the vital need of early intervention
because it recognises the lifelong impact of ASD on
language development, speech, cognitive, and social
skills, especially when symptoms appear during the first
two years of life. Respondents to nominal questions are
asked to provide a score that indicates the probability of
Autism Spectrum Disorder. Using toddler datasets, our
study demonstrates the efficacy of logistic regression in
producing precise predictions with little characteristics.
The study contributes to the larger objective of
improving the diagnostic process by highlighting the
importance of early discovery in reducing the long-term
impacts of ASD. Crucially, this method is presented as a
quick and affordable substitute for clinical testing,
providing an invaluable tool for enhancing diagnostic
accuracy in cases with toddler ASD.