ENBIS European Network for Business and Industrial Statistics
Sprache des Tagungstitel:
Biclustering of gene expression data, that is, clustering genes and
samples simultaneously, is an important unsupervised approach
to analyze transcriptomic data.
We introduce a novel generative model for biclustering
called ``Factor Analysis for Bicluster Acquisition'' (FABIA).
FABIA is exploratory factor analysis where both the factors and
the loadings are sparse, that is they contain many zeros.
For each factor, the posterior factor values and the factor's
loading vector determine the membership of samples and genes,
respectively, to the bicluster associated with this factor.
The degree of sparseness governs the size of the biclusters.
We report how FABIA performs on different gene expression data sets