Scholarlink

<p>The analysis of microarray data plays a critical role in cancer diagnosis, prognosis, and biomarker discovery, yet remains challenging due to high dimensionality and limited sample sizes. This study introduces the Multi-Objective Crayfish Optimization Algorithm (MOCOA), a multi-objective feature selection approach designed to efficiently explore the search space and identify informative gene subsets by simultaneously optimizing classification performance and feature reduction. To enhance convergence efficiency, MOCOA incorporates a mutual information-based initialization strategy that guides the search toward promising regions of the solution space. The proposed method is evaluated on four widely used microarray cancer datasets for cancer subtype classification and compared against established multi-objective optimization techniques, including Multi-Objective Particle Swarm Optimization (MOPSO), Multi-Objective Genetic Algorithm (MOGA), and Multi-Objective Grey Wolf Optimizer (MOGWO). Experimental results, assessed using the Inverted Generational Distance (IGD) metric, indicate that MOCOA achieves competitive and consistently improved Pareto front approximations across the evaluated datasets. These findings suggest that MOCOA provides an effective and robust framework for feature selection in microarray-based cancer classification tasks.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Mocoa-fs: a multi-objective feature selection approach for genetic cancer classification using microarray data

  • 期刊论文

  • 发表日期：

    2026年3月30日

查看全文
  • Jayashree Piri,
  • Himanshu Sekhar Panigrahi,
  • Raghunath Dey,
  • Andreas Kanavos,
  • Vassilis C. Gerogiannis,
  • Biswaranjan Acharya

摘要

The analysis of microarray data plays a critical role in cancer diagnosis, prognosis, and biomarker discovery, yet remains challenging due to high dimensionality and limited sample sizes. This study introduces the Multi-Objective Crayfish Optimization Algorithm (MOCOA), a multi-objective feature selection approach designed to efficiently explore the search space and identify informative gene subsets by simultaneously optimizing classification performance and feature reduction. To enhance convergence efficiency, MOCOA incorporates a mutual information-based initialization strategy that guides the search toward promising regions of the solution space. The proposed method is evaluated on four widely used microarray cancer datasets for cancer subtype classification and compared against established multi-objective optimization techniques, including Multi-Objective Particle Swarm Optimization (MOPSO), Multi-Objective Genetic Algorithm (MOGA), and Multi-Objective Grey Wolf Optimizer (MOGWO). Experimental results, assessed using the Inverted Generational Distance (IGD) metric, indicate that MOCOA achieves competitive and consistently improved Pareto front approximations across the evaluated datasets. These findings suggest that MOCOA provides an effective and robust framework for feature selection in microarray-based cancer classification tasks.