Subtype identification from heterogeneous TCGA datasets on a genomic scale by multi-view clustering with enhanced consensus

Table 3 The average NMIs and ACCs and standard errors obtained by the ECMC and other comparison partners on real benchmark datasets

	Methods	Cora	Texas	Wisconsin	Washington	Cornell
NMI	SV1	0.021 ±0.001	0.175 ±0.001	0.273 ±0.004	0.252 ±0.001	0.182 ±0.002
	SV2	0.004 ±0.000	0.098 ±0.001	0.064 ±0.001	0.096 ±0.002	0.083 ±0.001
	Concat	0.002 ±0.000	0.120 ±0.001	0.120 ±0.001	0.128 ±0.001	0.156 ±0.001
	Coreg	0.025 ±0.001	0.234 ±0.002	0.284 ±0.005	0.306 ±0.002	0.213 ±0.005
	SNF	0.013 ±0.000	0.156 ±0.003	0.303 ±0.001	0.204 ±0.006	0.200 ±0.001
	CMC	0.085 ±0.003	0.316 ±0.002	0.343 ±0.003	0.328 ±0.002	0.326 ±0.002
	ECMC	0.688 ±0.000	0.348 ±0.002	0.419 ±0.003	0.380 ±0.001	0.343 ±0.005
ACC	SV1	0.587 ±0.003	0.570 ±0.001	0.533 ±0.004	0.440 ±0.001	0.456 ±0.004
	SV2	0.544 ±0.000	0.563 ±0.001	0.462 ±0.002	0.490 ±0.001	0.453 ±0.001
	Concat	0.511 ±0.001	0.383 ±0.003	0.375 ±0.003	0.375 ±0.002	0.411 ±0.001
	Coreg	0.590 ±0.001	0.612 ±0.001	0.558 ±0.004	0.519 ±0.003	0.496 ±0.005
	SNF	0.549 ±0.000	0.601 ±0.000	0.587 ±0.003	0.551 ±0.006	0.497 ±0.000
	CMC	0.665 ±0.004	0.468 ±0.003	0.578 ±0.005	0.492 ±0.002	0.479 ±0.002
	ECMC	0.935 ±0.000	0.566 ±0.001	0.635 ±0.001	0.648 ±0.002	0.539 ±0.002

ISSN: 1755-8794