Figure 11

Significance of prominence values. Prominence values were assigned by the Eigenvector Centrality model applied to the focused subgraph with the protein query cancer. Distributions are shown for the first, second and third eigenspaces. The distributions of low-scoring nodes are modeled by normal distributions. Of the 1977 nodes in the focused subgraph, exactly 264 are more than 5 standard deviations apart from the mean of the normal distribution, in the first and second eigenspaces, and 44 in the third eigenspace. Note that these numbers correspond to the size of the connected components that are mapped to these eigenspaces (see section 'Eigenspaces and connected components'). The top scoring nodes in both the first and second eigenspaces are drawn from the largest CC, but different groups of documents are assigned the most significant values in each eigenspace.