Predicting ASD Using Optimized Machine Learning

Autistic Spectrum Disorder (ASD) is a neurodevelopmental disorder that affects the development of cognitive abilities, social-communicative skills and shows restricted or repetitive behaviors and interests. An earlier diagnosis of ASD reassures those specialized interventions that optimize children's long-term outcomes are accessed and met. Using a machine learning prediction model that can predict whether an individual has ASD at earlier ages can have multiple advantages. This paper Analyzes ASD datasets for children, adolescents and adults using NBC, KNN and LR. A 10-fold cross-validation method is used to evaluate the models with the three datasets. After that, the Attribute selection method is used to cut down the features using Info gain evaluator. Lastly, two ML models are stacked using the vote stacking technique to be evaluated. This paper is done with the aim to optimize predicting ASD using Autistic-Spectrum Quotient.