A new fangled approach for grading autism machine learning and deep learning techniques

Abstract

Autistic Spectrum Disorder (ASD) is primarily related to genetic newlineand neurological entities resulting in challenges faced in social interaction and newlinecommunication. As per WHO statistics, the number of patients diagnosed with newlineASD has seen a slow rise. Premature diagnosis of ASD with pre-planned newlinetreatment would aid the child to get out of the spectrum and have a life as newlineusual. The treatment planning primarily relies on paying attention to the newlinedevelopmental regions that lag in the children. ASD begins with a newlinedevelopmental delay that tends to become serious if the right treatment is not newlinegiven at the premature stage. Several recent studies highlight clinical newlinediagnosis, therapy monitoring, and brain image analysis, but they are not newlineattentive towards the diagnosis of ASD with important treatment area newlinedetection depending on machine learning and deep learning. The objective of newlinethe work is to categorize the ASD data to render a rapid, accessible, and newlinesimple means of supporting the early ASD diagnosis with their primary newlinespecification of the treatment area. Nearly all the research was dependent on newlineCARS, ADOS, ABIDE datasets. In this, the ISAA dataset is utilized for newlineclassifying the ASD level and domains in the ISAA scale are used for finding newlinethe lagging regions of the patient for further treatment. newlineThe first contribution is involved with the pre-processing of the newlinedataset to eliminate the Null Values, Redundant Values, and Missing Values. newlineFeature extraction and further feature selection are carried out with the help of newlineParticle Swarm Optimization. Later, the Improved Adaptive Neuro-Fuzzy newlineInterference System (IANFIS) classification algorithm is used for diagnosing newlinethe autism level with the lagging areas for better treatment. newline