International Journal of applied mathematics and computer science

online read us now

Paper details

Number 4 - December 2024
Volume 34 - 2024

A novel chaotic binary butterfly optimization algorithm based feature selection model for classification of autism spectrum disorder

Anandkumar Ramakrishnan, Rajakumar Ramalingam, Padmanaban Ramalingam, Vinayakumar Ravi, Tahani Jaser Alahmadi, Siti Sarah Maidin

Abstract
Autism spectrum disorder (ASD) issues formidable challenges in early diagnosis and intervention, requiring efficient methods for identification and treatment. By utilizing machine learning, the risk of ASD can be accurately and promptly evaluated, thereby optimizing the analysis and expediting treatment access. However, accessing high dimensional data degrades the classifier performance. In this regard, feature selection is considered an important process that enhances the classifier results. In this paper, a chaotic binary butterfly optimization algorithm based feature selection and data classification (CBBOAFS-DC) technique is proposed. It involves, preprocessing and feature selection along with data classification. Besides, a binary variant of the chaotic BOA (CBOA) is presented to choose an optimal set of a features. In addition, the CBBOAFS-DC technique employs bacterial colony optimization with a stacked sparse auto-encoder (BCO-SSAE) model for data classification. This model makes use of the BCO algorithm to optimally adjust the ‘weight’ and ‘bias’ parameters of the SSAE model to improve classification accuracy. Experiments show that the proposed scheme offers better results than benchmarked methods.

Keywords
data classification, feature selection, metaheuristics, machine learning, autism spectrum disorder

DOI
10.61822/amcs-2024-0043