A Survey of Machine Learning Techniques Related to Understanding Autism Spectrum Disorder

Speech, social skills, non-verbal communication, and repetitive activities are all indicators of autism spectrum disorder (ASD), a disease classified as a spectrum disorder by the DSM-5. Gender plays a role in the prevalence of ASD. If ASD isn't caught early enough, the damage is done. It's been a while since computer-aided ASD diagnosis made a lot of progress, which is talked about in this work. This article explores virtual reality for people with ASD with a focus on training and intervention. Our taxonomy categorizes prior VR training for people with ASD into two categories: immersive and regular (non-immersive) VR systems, which we present and put into practice. To fill up the knowledge gaps, we conducted a comprehensive analysis of the existing literature and offered suggestions for further investigation. It is the goal of this study to categorize and review various machine learning approaches, as well as analyze the characteristics of ASD and performance evaluation using various measures. People with ASD have difficulties with social interaction, communication, and other aspects of their cognitive and behavioural development. According to this review, the Random Forest method outperforms other algorithms.