Chaotic emperor penguin optimised extreme learning machine for microarray cancer classification.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Microarray technology is crucial for analyzing numerous genes and samples in cancer classification.
• Efficient analysis of complex microarray data necessitates advanced intelligent techniques.

Purpose of the Study:

• To propose a novel hybrid technique for enhanced cancer classification using microarray data.
• To improve the accuracy and efficiency of gene selection and classification models.

Main Methods:

• A hybrid approach combining Fisher criterion and ReliefF for initial gene selection.
• Utilizing a chaotic emperor penguin optimization (CEPO) algorithm to pre-train an extreme learning machine (ELM) by optimizing weights and biases.
• Independent filtering using Fisher score and ReliefF for relevant gene identification.

Main Results:

• Experiments conducted on seven benchmark datasets demonstrated superior performance.
• The proposed CEPO-ELM hybrid method achieved higher accuracy compared to existing state-of-the-art techniques.
• The method's effectiveness was validated through metrics including accuracy, sensitivity, specificity, and F-measure.

Conclusions:

• The novel hybrid technique offers a significant advancement in cancer classification using microarray data.
• The integration of CEPO for ELM pre-training enhances classification performance.
• The proposed method provides a robust and accurate framework for analyzing high-dimensional genomic data.