MORPSO_ECD+ELM: A Unified Framework for Gene Selection and Cancer Classification

Area of Science:

• Bioinformatics
• Computational Biology
• Machine Learning

Background:

• Gene selection and cancer classification are complex multi-objective problems.
• Current methods often fail to balance competing objectives like accuracy and gene relevance.
• Independent optimization limits the effectiveness of gene selection and classification.

Purpose of the Study:

• To propose a unified framework, MORPSO_ECD+ELM, for simultaneous multi-objective optimization of gene selection and cancer classification.
• To enhance diversity preservation and solution space exploration in multi-objective optimization.
• To achieve robust and efficient cancer classification with biologically meaningful gene identification.

Main Methods:

• Formulating gene selection and classification as a multimodal multi-objective optimization problem (MMOP).
• Introducing an enhanced crowding distance (ECD) metric for improved diversity.
• Utilizing an advanced multi-objective particle swarm optimization variant (MORPSO_ECD) with ring topography.
• Integrating the framework with the Extreme Learning Machine (ELM) for classification.

Main Results:

• The MORPSO_ECD+ELM framework effectively optimizes classification accuracy and gene selection simultaneously.
• Experimental validation demonstrates high classification performance.
• The method successfully identifies biologically relevant gene subsets.
• The approach provides a powerful solution for integrating gene selection and cancer classification.

Conclusions:

• The proposed unified framework offers a significant advancement in addressing multi-objective challenges in cancer research.
• MORPSO_ECD+ELM achieves superior performance in both classification accuracy and the identification of meaningful biomarkers.
• This integrated approach bridges the gap between computational methods and biological interpretability in cancer genomics.