Feature Selection in Breast Cancer Gene Expression Data Using KAO and AOA with SVM Classification

Area of Science:

• Bioinformatics
• Computational Biology
• Machine Learning in Healthcare

Background:

• Breast cancer classification from gene expression data is challenging due to high dimensionality.
• Existing methods struggle with feature redundancy and premature convergence.

Purpose of the Study:

• To develop a novel hybrid optimization framework for effective gene selection and breast cancer classification.
• To improve the accuracy and efficiency of cancer prediction using gene expression profiles.

Main Methods:

• A hybrid framework combining Kashmiri Apple Optimization Algorithm (KAO) and Armadillo Optimization Algorithm (AOA) for feature selection.
• Support Vector Machines (SVM) employed for precise classification of breast cancer subtypes.
• A dual-stage optimization approach for global gene exploration and local refinement.

Main Results:

• Achieved 98.97% classification accuracy, 98.46% precision, 100% recall, and 99.22% F1-score.
• Identified a highly informative subset of 15 genes for accurate breast cancer classification.
• Demonstrated robust and consistent high performance across varying gene subset sizes.

Conclusions:

• The KAO-AOA hybrid framework offers a powerful tool for gene-based cancer prediction.
• Optimized feature selection reduces redundancy and enhances classification accuracy.
• The framework shows potential for application to diverse cancer datasets and clinical settings.