Joint Transfer Extreme Learning Machine with Cross-Domain Mean Approximation and Output Weight Alignment

Table 3

Accuracy of different algorithms on MSRC + VOC2007 and Reuters-21578 datasets.

Methods\dataset	Nontransfer learning algorithm				Transfer learning algorithm
	1NN	SVM	ELM	SSELM	TCA1	TCA2	JDA1	JDA2	DAELM_S	DAELM_T	ARRLS	TELM-OWA	JTELM
MSRC vs. VOC	35.95	35.10	38.56	37.52	27.52	27.12	35.82	37.52	38.34	33.60	34.64	41.37	40.32
VOC vs. MSRC	45.23	54.69	58.00	63.20	47.44	48.31	54.22	54.61	58.37	30.61	50.35	66.22	70.76
Average	40.59	44.90	48.28	50.36	37.48	37.71	45.02	46.06	48.35	32.10	42.50	53.79	55.54
Orgs vs. people (1)	72.85	75.25	77.24	80.30	74.09	73.76	75.99	80.63	76.64	48.71	80.55	63.74	83.28
People vs. orgs (2)	72.03	77.12	79.30	84.72	73.65	74.54	77.12	85.85	78.41	47.72	84.56	64.27	85.53
Orgs vs. place (3)	67.50	70.18	63.57	69.61	70.66	71.72	70.95	74.98	68.53	43.81	72.20	68.09	76.32
Place vs. orgs (4)	61.12	63.77	64.57	71.16	68.80	70.18	67.52	75.98	68.77	42.50	74.11	74.68	78.74
People vs. place (5)	52.65	60.63	57.01	66.02	61.84	62.58	59.89	64.07	59.14	42.46	65.18	59.04	67.5
Place vs. people (6)	53.39	57.94	58.68	61.65	60.82	60.82	54.60	57.38	58.77	39.93	58.56	60.45	64.34
Average	63.26	67.48	66.73	72.24	68.31	68.93	67.68	73.15	68.38	44.19	72.53	65.05	75.95
Total average	57.59	61.84	62.12	66.77	60.60	61.13	62.01	66.38	63.37	41.17	65.02	62.23	70.85

Bold values indicate that the value is the best result of the row in which it is located.