Autism spectrum disorder classification using Adam war strategy optimization enabled deep belief network

Abstract

Autism spectrum disorder (ASD) is a brain disorder caused by dysfunction in the brain. ASD patients have social interaction and communication problems that are determined by numerous deep learning (DL) methods. The existing ASD detection methods are complex and inaccurate in ASD classification. In this investigation, the patient with ASD is determined by the Adam war strategy optimization (AWSO) based Deep Belief Network (DBN). The developed AWSO algorithm is modelled by assimilating the Adam optimizer with the War Strategy Optimization (WAO). The Adam war strategy optimization technique is a simple process and it overcomes the issue of the existing method with outstanding performance. The pre-processing is finished using anisotropic diffusion and Region of Interest (ROI) extraction to remove the noise in the input images. testing Moreover, the pivotal region extraction is completed by the Box Neighbourhood Search Algorithm based on Functional Connectivity to progress the performance of ASD classification. Then, the ASD classification is completed by the DBN, and the AWSO algorithm establishes the learning of DBN. Here, the analysis of AWSO-DBN is done using ABIDE-I and ABIDE-II, and the AWSO-DBN attained outstanding performance with the ABIDE-I dataset by varying the training set. The experimental outcome reveals that the AWSO-DBN algorithm achieved a better specificity of 0.935, an accuracy of 0.924, and a sensitivity of 0.930.