Abstract:

     The degree of difficulty in finding significant biclusters in gene expression data grows with the increase of the size of the dataset. This paper presents an intelligent biclustering method which makes full use of the climbing ability of tabu search and genetic algorithm, balances relation between regional search and global search, and makes the results independent of a pre-defined threshold. Experiments show that the method has a better convergence property and can discover better biclusters than traditional methods.

Key words: genetic algorithm , tabu search , biclustering , gene expression data