Adaptive Correlation-Based Gene Expression Analysis Using Enhanced Ensemble Biclustering Framework

The analysis of gene expression data plays a pivotal role in understanding complex biological functions, disease mechanisms, and gene regulation patterns. While biclustering methods such as Bimax have improved the identification of local gene-condition patterns, limitations persist in terms of computational complexity and biological relevance. This study proposes an Adaptive Ensemble Biclustering Framework (AEBF) that integrates multiple correlation-based biclustering algorithms, including an improved version of the modified Bimax, to enhance gene expression pattern discovery. The framework incorporates dynamic z-score normalization, adaptive outlier detection, and ensemble scoring based on size, coherence, and biological enrichment. Experimental validation on benchmark microarray datasets reveals that AEBF not only improves bicluster consistency and biological relevance but also demonstrates significant computational efficiency compared to standalone biclustering methods.