Enhanced Whale Optimization and Ensemble Learning Algorithm for Autism Spectrum Disorder Classification

Autism spectrum disorder (ASD) is an array of neurodevelopmental illnesses characterized by difficulties with sociability and interaction and restricted repetitive activities.The amount of people with ASD has surged in recently, and the underlying etiology of the condition is still unknown.Although ASD cannot be cured, early discovery is ideal since it makes mitigation care more effective.Given what extent the signs are, an ASD diagnoses may give in the future, although it usually occurs around the age of two.Classification accuracy and feature selection are not substantially guaranteed in the current approach.To address these problems, Enhanced Whale Optimization Algorithm (EWOA) and Ensemble Modified Auto Encoder Convolution Neural Networks (EM-AECNN)-Artificial Neural Network (ANN) system are suggested for ASD classification n excellently.The suggested study contains pre-processing, feature selection, and classification.This work contains three main modules such as pre-processing, feature selection and categorization.Data pre-processing is a technique is executed through Synthetic Minority Oversampling Technique (SMOTE)which is utilized by eliminating the ASD dataset's unnecessary data.The EWOA method is utilized to choose the pertinent characteristics from the autism dataset during the feature selection process.By choosing the optimal fitness features, the objective function of EWOA is employed to increase classification accuracy.In order to categorize the accurate ASD outcomes, the EM-AECNN-ANN technique is ultimately employed.Given the testing outcome, it is determined that the suggested EM-AECNN-ANN algorithm provides greater precision, accuracy, recall and f-measure than the current approaches.