Unsupervised Domain Adaptation Using Exemplar-SVMs with Adaptation Regularization

Table 2

We also conduct our experiments for the tasks of multidomain and gain an improvement comparing with methods proposed before. The experiments adopt the same strategy as the single domain adaptation. We treat multidomain as one source or target to find the shared features in a latent space. However, the complexity of the multidomain shared features limits the accuracy of tasks [average standard error of accuracy (%)].

Task	SVM	KMM	TCA	TJM	SA	GFK	LSSA	LRESVM	DAESVM
	45.7	37.4	40.5	57.1	59.4	47.3	61.7	80.1
	37.1	31.6	43.0	60.2	48.7	47.6	74.2	86.9
	41.4	43.8	57.2	63.9	51.9	51.4	77.0	82.9
	43.9	50.6	54.9	69.0	60.2	60.4	63.7	87.7
	71.0	61.0	54.0	61.3	54.0	47.0	71.9	80.8
	81.4	53.9	77.4	71.8	57.4	64.1	80.7	89.3
Average	53.4	46.4	54.5	63.9	55.2	53.0	71.5	84.6