Input &nbsp;&nbsp;&nbsp; Input Microarray Gene Expression Data

• m × n gene expression matrix• Rows: genes• Columns: conditions / samples

Projection into Column-Pair Spaces

• Generate all column-pairs• Map genes into 2-D spaces

Geometric Pattern Detection using Hough Transform

• Line detection in column-pair spaces• Identify additive & multiplicative patterns• Extract sub-biclusters

Sub-Bicluster Representation

• Binary encoding of sub-biclusters• Grouping based on shared columns

Hypergraph Construction

• Vertices: sub-biclusters• Hyperedges: shared gene-condition relationships

Hypergraph Partitioning & Optimization

• Spectral / Hmetis partitioning• Remove weak sub-bicluster relationships

Feature Extraction from Optimized Biclusters

• Select coherent gene subsets• Dimensionality reduction

Random Forest Classification

• Trained on optimized biclusters• Cancer classification & prediction

Output

• Improved classification accuracy• Biologically interpretable biclusters

asfdsdf