EGFAFS: A Novel Feature Selection Algorithm Based on Explosion Gravitation Field Algorithm.

Area of Science:

• Bioinformatics
• Machine Learning
• Data Mining

Background:

• High-dimensional feature spaces in gene expression data pose challenges for traditional analysis methods.
• Feature selection (FS) is crucial for effective data mining and machine learning with complex datasets.
• Existing methods may struggle with the scale and complexity of gene expression data.

Purpose of the Study:

• To introduce a novel feature selection algorithm, EGFAFS (Explosion Gravitation Field Algorithm for Feature Selection).
• To enhance the efficiency and accuracy of feature selection specifically for gene expression data.
• To validate the performance of EGFAFS against established feature selection techniques.

Main Methods:

• Developed EGFAFS, integrating the Explosion Gravitation Field Algorithm.
• Constructed a feature pool using Random Forests and the Gini index to prioritize relevant features.
• Evaluated EGFAFS on eight gene expression datasets, comparing it with eight other FS methods.

Main Results:

• EGFAFS demonstrated superior performance in feature selection on gene expression datasets compared to eight other algorithms.
• The algorithm achieved better results across various evaluation metrics.
• Genes identified by EGFAFS were found to be significant in differential co-expression networks and biological functions.

Conclusions:

• EGFAFS is a highly effective algorithm for feature selection in high-dimensional gene expression data.
• The method offers improved efficiency and accuracy, contributing to better biological insights.
• EGFAFS represents a significant advancement for analyzing complex biological datasets.