Efficient Hybrid algorithm for Gene Selection Based on Chaos Game Optimization

Gene selection stands as a crucial preprocessing step in biomedical data mining and holdssignificance in various biological applications, particularly in biomarker identification. Despite theintroduction of numerous gene selection algorithms, challenges persist, including prolongedconvergence time, parameter tuning, and suboptimal performance. This paper introduces an efficientapproach employing the Chaos Game Optimization (CGO) algorithm for the selection of relevant genesfrom large-scale gene datasets. Our algorithm operates in two distinct phases: during the Relief-basedfiltering phase, weights are assigned to genes, and in the subsequent CGO-based wrapping phase,optimal genes are identified. To assess the efficacy of our proposed method, we conduct evaluationsusing standard gene expression datasets. The experimental results highlight that, in comparison toanalogous methods, our approach produces a more streamlined set of features without compromisingaccuracy.

کلیدواژه ها:

Gene selection ، Chaos Game Optimization ، High dimensionality ، Bioinformatics