Training Classifiers under Covariate Shift by Constructing the Maximum Consistent Distribution Subset

Table 2

The running time and parameter values in Experiment  1.

Data set	SVM	IDC1	TrAdaBoost
	Australian	250.6, 10, 16	210.1, 100, 16	280.9, 50, 128, 100
Balance	322.1, 20, 8	295.2, 50, 32	353.2, 30, 256, 100
Breast	412.1, 100, 64	368.8, 10, 128	515.5, 100, 32, 100
Cleveland	243.2, 50, 32	205.1, 20, 64	301.5, 100, 64, 100
Credit	1129.2, 1000, 128	989.3, 500, 16	1250.6, 500, 64, 100
Diabetes	517.6, 1000, 32	410.1, 100, 128	620.3, 500, 32, 100
Heart	32.5, 100, 256	24.6, 30, 512	33.7, 10, 128, 100
Ionosphere	50.6, 20, 128	42.1, 1000, 32	52.3, 10, 256, 100
Iris	12.2, 10, 64	9.3, 50, 64	13.1, 100, 32, 100
Liver	19.6, 30, 32	18.2, 50, 32	22.3, 100, 128, 100
Page	22.1, 100, 128	18.8, 200, 64	22.8, 500, 16, 100
Sonar	9.6, 50, 1024	6.6, 100, 128	10.2, 80, 512, 100
Thyroid	32.3, 1000, 256	26.6, 500, 128	40.2, 60, 64, 100
Vehicle	1080.9, 100, 8	916.1, 50, 16	1120.3, 100, 128, 100
Voting	62.8, 60, 64	56.6, 100, 256	70.5, 80, 512, 100
Waveform21	632.1, 30, 32	527.1, 60, 64	636.9, 10, 16, 100
Waveform40	535.3, 50, 64	468.2, 100, 128	586.2, 200, 16, 100
Wine	50.2, 1000, 256	44.1, 500, 128	59.2, 200, 1024, 100
wdbc	289.2, 200, 128	266.1, 100, 256	299.5, 500, 32, 100
wpbc	267.1, 500, 512	256.9, 1000, 1024	268.2, 100, 512, 100