Feature Selection by Hybrid Brain Storm Optimization Algorithm for COVID-19 Classification.

Area of Science:

• Computer Science
• Machine Learning
• Data Science

Background:

• High-dimensional data presents challenges in machine learning due to increased noise and computational complexity.
• Effective feature selection is crucial for improving predictive model performance and reducing processing time.
• Existing methods often struggle to balance feature relevance and dimensionality reduction.

Purpose of the Study:

• To propose a novel binary hybrid metaheuristic algorithm for optimal feature subset selection.
• To enhance the performance of machine learning models by reducing data dimensionality and noise.
• To validate the algorithm's effectiveness on various classification datasets, including a coronavirus disease dataset.

Main Methods:

• A hybrid metaheuristic approach combining the Brain Storm Optimization (BSO) algorithm with the Firefly Algorithm (FA).
• The hybrid algorithm is implemented as a wrapper method for feature selection in classification tasks.
• Evaluation involved testing on 21 diverse datasets and comparison against 11 existing metaheuristic algorithms.

Main Results:

• The proposed hybrid algorithm demonstrated robust performance in reducing and selecting optimal feature subsets.
• Experimental results showed significantly higher classification accuracy compared to other state-of-the-art methods.
• The method proved effective in handling high-dimensional data and improving model generalization.

Conclusions:

• The developed binary hybrid metaheuristic algorithm offers a superior approach to feature selection for classification problems.
• This method provides an efficient and accurate solution for dimensionality reduction in machine learning.
• The algorithm's robustness and effectiveness are confirmed by its performance across multiple datasets and its application to real-world health data.