Elitist Binary Wolf Search Algorithm for Heuristic Feature Selection in High-Dimensional Bioinformatics Datasets.

Area of Science:

• Bioinformatics
• Machine Learning
• Computational Biology

Background:

• High-dimensional datasets pose significant challenges for accurate classification.
• Traditional feature selection methods often struggle with complexity and computational cost.

Purpose of the Study:

• To develop an optimized feature selection method using the Wolf Search Algorithm (WSA) for high-dimensional datasets.
• To enhance classification accuracy and reduce computational time in bioinformatics.

Main Methods:

• A novel Wolf Search Algorithm (WSA) incorporating memory and binary strategies for feature selection.
• A wrapper strategy combining the optimized WSA with an extreme learning machine classifier.
• Evaluation on six public high-dimensional bioinformatics datasets.

Main Results:

• The proposed WSA method achieved superior classification accuracy, outperforming conventional methods by up to 29%.
• Demonstrated significant improvements in computational efficiency, exceeding previous WSAs by up to 99.81%.
• Successfully identified optimal feature subsets for improved global classification model performance.

Conclusions:

• The proposed WSA-based feature selection method is highly effective for high-dimensional bioinformatics data.
• Offers a computationally efficient and accurate alternative to existing feature selection techniques.
• Highlights the adaptability of evolutionary algorithms in solving complex machine learning problems.