Fig. 3.

The bicluster model experiment on the synthetic datasets. (A) The relevance score of bicluster methods on synthetic gene expression data [7] with different noise level; (B) The relevance score of bicluster methods on synthetic gene expression data [7] of the different number of biclusters; (C) The relevance score of bicluster methods on synthetic spectroscopic data 1 (Fig. 2) with different noise level; (D) The average relevance score of bicluster methods on synthetic spectroscopic data 1 (Fig. 2) the different number of biclusters. The methods: NMF (red circle), PCA (blue circle), Sparse NMF(green circle), BiMax(red triangle), Plaid(green triangle), Spectral(yellow triangle), Xmotifs(blue triangle), Fabia(magenta triangle) and ISA(cyan triangle. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)