Hubs of knowledge: using the functional link structure in Biozon to mine for biologically significant entities

BMC Bioinformatics

Table 2 Comparison of prominence models on the global Biozon graph.

Query Term	Query Type	Prominence Model	Average No. Neighbors	Average Consistent Neighbors	Average Ratio	Q(R)	UROC(R)
		Hubs & Authorities	4.38	0.6	0.13	167	4879
ubiquitin	protein	Eigenvector Centrality	4.38	0.6	0.13	167	4879
		Katz's Status	3.82	0.48	0.12	135	3436
		PageRank	12.18	3.56	0.27	977	26021
		Hubs & Authorities	1.19	0.79	0.48	357	10086
stromelysin	protein	Eigenvector Centrality	1.19	0.79	0.48	357	9912
		Katz's Status	1.19	0.79	0.48	357	10061
		PageRank	1.19	0.79	0.48	357	11593
		Hubs & Authorities	5.22	1.08	0.25	317	6137
cancer	protein	Eigenvector Centrality	5.20	0.9	0.21	245	5346
		Katz's Status	4.84	0.64	0.19	170	4915
		PageRank	6.68	1.8	0.30	535	16082
		Hubs & Authorities	1.26	0.5	0.46	166	4628
cancer	nucleic	Eigenvector Centrality	1.26	0.46	0.45	151	4635
		Katz's Status	1.08	0.44	0.44	146	4529
		PageRank	1.60	0.52	0.39	167	4633
		Hubs & Authorities	1.1	0.67	0.58	223	4520
autoimmune	protein	Eigenvector Centrality	1.1	0.67	0.58	223	4520
		Katz's Status	1.1	0.67	0.58	223	4520
		PageRank	1.1	0.67	0.58	223	4582
		Hubs & Authorities	0.98	0.4	0.4	226	4559
autoimmune	nucleic	Eigenvector Centrality	0.98	0.4	0.4	226	5430
		Katz's Status	0.98	0.4	0.4	226	4401
		PageRank	0.98	0.4	0.4	226	7659

Only the principal eigenspace is used. For each query consisting of (query term, query type, prominence model) we report the following results over the set R of the top 50 documents returned by that query: The average number of neighbors per document in that set, the average number of consistent neighbors, the average ratio of these numbers, the quality of results set Q(R), and the most informative measure UROC(R).

ISSN: 1471-2105