Research Article

Discriminative Extreme Learning Machine with Cross-Domain Mean Approximation for Unsupervised Domain Adaptation

Table 4

Classification accuracies (%) of all algorithms on MSRC + VOC2007 and reuters-21578.

Data setMethods
1NNSVMELMSSELMTCA1TCA2JDA1JDA2DAELM_SDAELM_TARRLSTELM-OWAELM-CDMADELM-CDMADKELM-CDMA

MSRC ⟶ VOC36.035.138.637.527.934.927.134.838.333.634.641.438.038.239.0
VOC ⟶ MSRC45.254.758.063.247.954.248.354.658.430.650.466.262.763.564.5
Orgs vs people (1)72.975.377.280.376.577.580.181.876.648.780.663.782.583.482.2
People vs orgs (2)72.077.179.384.777.379.485.185.378.447.784.664.386.186.385.4
Orgs vs place (3)67.570.263.669.672.471.875.776.268.543.872.268.178.379.178.2
Place vs orgs (4)61.163.864.671.268.265.876.476.168.842.574.174.779.580.683.3
People vs place (5)52.760.657.066.061.260.666.965.859.142.565.259.068.969.367.2
Place vs people (6)53.457.958.761.756.151.158.151.258.839.958.660.566.565.169.0
Total average57.661.862.166.860.961.964.765.763.441.265.062.270.370.771.1