Adaptive Ensemble with Human Memorizing Characteristics for Data Stream Mining

Table 4

Results of average chunk training time 10⁻³ second.

Dataset	Win	SEA	AWE	ACE	MAE
SEA	1.43	6.93	4.44	14.99	14.75
Tree	4.86	10.95	8.42	37.01	19.38
Hyp	6.39	12.24	9.67	42.88	20.70
RBF	16.50	22.23	19.71	77.93	30.80
LED	5.74	12.68	9.56	23.81	22.30
Adult	2.79	6.06	5.73	28.96	10.11
Bank	1.86	7.96	5.81	13.84	10.95
Conn	4.56	9.11	7.29	21.05	14.04
Cover	6.96	12.62	10.14	80.92	20.56
EEG	2.41	5.11	3.97	13.76	7.85
Elec	2.10	7.08	4.89	20.02	11.15
Page	3.02	7.15	4.78	12.09	4.73
Poker	2.19	7.90	5.40	48.55	16.38
Person	2.63	9.87	7.61	15.25	12.33
Robot	9.82	13.01	11.24	20.72	13.54
Mean	4.88	10.06	7.91	31.45	15.30