Semisupervised Community Preserving Network Embedding with Pairwise Constraints

Table 3

Accuracy (%) of node clustering.

Methods	Polbooks	Cornell	Texas	Washington	Wisconsin	Email	Citeseer	Cora
DeepWalk	52.29	33.29	40.44	39.12	35.26	30.18	48.01	20.64
Line	41.33	32.24	33.83	32.61	33.55	24.14	24.30	23.44
node2vec	67.77	32.62	33.81	34.37	31.79	21.97	23.99	26.95
GraRep	42.86	35.15	38.98	37.67	38.22	49.81	43.16	50.94
MNMF	71.91	39.73	45.96	52.09	43.49	50.41	27.68	35.60
NECS	70.38	45.41	55.03	50.37	47.21	14.08	25.75	29.62
DNR (1)	72.65	38.40	55.04	59.32	41.80	20.15	31.87	38.28
DNR (2)	73.08	36.97	48.38	50.39	43.73	17.49	40.17	31.68
DNR (5)	75.03	35.60	48.05	58.64	40.36	15.33	47.46	41.05
DNR (10)	72.63	33.63	39.75	43.06	47.42	16.34	34.77	29.53
SMNMF (1)	75.05	41.58	47.10	55.26	45.98	58.95	70.82	63.12
SMNMF (2)	78.38	43.82	51.04	57.21	48.45	65.85	75.85	65.16
SMNMF (5)	91.43	68.52	55.74	65.16	69.08	67.86	79.24	69.95
SMNMF (10)	100.00	76.61	72.19	71.35	74.78	70.93	82.12	74.12