ABSTRACT

Autism is one of the most complex and different types of disorders, and it necessitates precise diagnosis based on characteristics including societal engagement, limited verbal communication, and repetitive behaviour.A timely and correct diagnosis of Autism Spectrum Disorder can ensure that you receive the appropriate medical treatment and control to help you recover. In this publication, Artificial Neural Networks are used to classify anxiety-related arousal in children with Autism Spectrum Dis-order(ASD), which is one of the most important fields of data mining research.The main objective of this study is to apply the Artificial Neural Network Algorithm (ANN), Convolutional Neural Network Algorithm called as (CNN), in ASD using the Adaptive Kalman Filter Gaussian Mixture Model (AKFGMD). In this paper we have developed a proposed hybrid algorithm which was implemented in ASD dataset and the result relevant to the proposed algorithm is better result when compared with other classifications method with respect to accuracy, sensitivity and specificity.

Keywords: Adaptive Kalman Filter Gaussian Mixture Model (AKFGMD), Artificial Neural Network Algorithm, Convolutional Neural Network Algorithm, Hybrid CNN, Autism Spectrum Disorder