Classification of Children with/without Autism Spectrum Disorder Using Speech Signal

Autism Spectrum Disorder (ASD) is a neurodevelopmental and neurological disorder related to brain development, leading to problems of social communication and interaction. While there is no cure for ASD, effective and early interventions can improve its symptoms. Hence screening this problem from early ages is very important. In our research, speech of children and machine learning are used to classify children with ASD from typically developing ones. Obtained results show that the combination of speech features and k-Nearest Neighbor model is a promising approach for early detection of ASD.