Skip to main content
Fig. 1 | BMC Medical Genomics

Fig. 1

From: Comprehensive discovery of subsample gene expression components by information explanation: therapeutic implications in cancer

Fig. 1

A simple hierarchical model constructed by CorEx. Variables in each layer are associated with latent factors in the next higher layer with a weight that is determined by optimizing over its relative contribution to the explanatory latent factors. Visually we set the thickness of links based on mutual information between variables and the thickness of nodes is proportional to the information contributed by that factor toward the objective, TC(X;Y)

Back to article page