Fajila, Mohamed Nisper Fathima and Yusof, Yuhanis (2025) Multi-Leader dynamic composite salp swarm algorithm for biomarker selection in cancer classification. IAENG International Journal of Computer Science, 52 (10). pp. 3767-3776. ISSN 1819656X

Preview

PDF - Published Version Download (1MB) | Preview

Abstract

Biomarker selection for microarray-based cancer classification is a recent alternative in cancer diagnosis and prognosis. Though existing approaches have yielded competing performance in terms of classification accuracy and produce smaller gene subsets, a generalized classifier that can operate on various cancer microarray datasets is yet to be reported. Recently, metaheuristic algorithms have demonstrated momentous performance in cancer classification. Nevertheless, the performance of a metaheuristic algorithm is influenced by the fitness value, convergence, exploration, and exploitation capabilities. Thus, this study proposes a new variant of Salp Swarm Algorithm (SSA) which is later integrated with Correlation-based Feature Selection (CFS) filter for gene selection in cancer classification. The slow convergence issue in SSA is solved by introducing dynamic size solutions that are generated using a composite position update function. Further, the local optima issue is overcome by the population reinitialization method. Adding to that, exploitation of best solution search space is enhanced by multi-leaders which are termed as dual leaders. The proposed hybrid algorithm, named the CFS-Multi-Leader Dynamic Composite Salp Swarm Algorithm (CFS-MDCSSA), is evaluated using two metrics, namely classification accuracy and gene subset size, using Support Vector Machine (SVM) classifier. The proposed CFSMDCSSA-SVM achieved 100% accuracy with only a few biomarkers for all six cancer microarray datasets, reflecting the competitive performance of the proposed algorithm in gene selection for cancer classification.

Item Type:	Article
Uncontrolled Keywords:	Biomarker Selection, Cancer Classification, Microarray, Salp Swarm Algorithm
Subjects:	Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Divisions:	School of Computing
Depositing User:	Mrs. Norazmilah Yaakub
Date Deposited:	29 Oct 2026 14:45
Last Modified:	29 Oct 2026 14:45
URI:	https://repo.uum.edu.my/id/eprint/36646

Actions (login required)

View Item