Behavioral Features based Autism Spectrum Disorder Detection using Decision Trees

Abstract

            Autism Spectrum Disorder (ASD) could also be a neuro biological process condition affects a person’s noesis and behavior. the current diagnostic models area unit subjective and behavior dependent. The delay in identification at associate early age and makes it more durable to differentiate syndrome from alternative biological process disorders. the data mining ways area unit applied to urge the ASD with activity and neuro image parameters. The classification techniques area unit applied on the identification knowledge values.

Early identification and follow-up treatments makes a significant impact on unfit folks. The electrodes connected with the scalp area unit used to capture the graph signals. Abnormalities in encephalogram (EEG) area unit typically used as reliable biomarkers to diagnose ASD. The ASD classification is run with graph signal process and learning models. The graph channel frequency minimisation is combined inside the prediction mechanism. applied mathematics options area unit extracted from noise filtered graph knowledge before and once separate riffle remodel. Relevant options and graph channels were selected mistreatment correlation-based feature choice. the coaching models and have vectors area unit used to minimize the graph channels. The Naïve Bayes classifier is used inside the ASD prediction method.

The adult syndrome discovery model is supposed to analysis the activity options of the patients on top of eighteen years. The activity options area unit collected with the support of a verbaliser. the selection tree classifier is employed to urge the illness levels. The syndrome discovery method is compared with the Naïve Bayes and Support Vector Machine (SVM) classification techniques