Autistic Spectrum Disorder Screening Classification with Machine Learning Approaches

The degree of severity and range of symptoms of Autistic Spectrum Disorder (ASD) differ widely among young children, and their medical conditions are determined by their age, psychological features, speech ability, and accompanied disorders. Early detection of this disease can significantly reduce medical costs. However, the waiting time for a diagnosis of ASD is prolonged, and treatments are exorbitantly costly. In this era of big data, conventional techniques for disease classification and prediction have proven unsuccessful when compared to machine learning approaches. The current study compared the performance ofmultiple machine learning models, including Gaussian Naive Bayes (GNB), K- Nearest Neighbor (KNN), and Decision Tree (DT), in the classification of ASD traits. The study's findings should enable health stakeholders to emphasise the importance of early detection, which will be beneficial in the long run for the management of this health condition. Overall, the findings of this study will help parents, healthcare professionals, and other interested parties support children with ASD and maintain medical stability. Following the exploratory comparison conducted for this study project, the KNN is more accurate (100%) in classifying ASD traits than the other two machine learning models.