Research Article
Efficient Multiple Kernel Learning Algorithms Using Low-Rank Representation
Table 1
The performances of MKL algorithms and LR-MKL algorithms on the datasets Yale, ORL, LSVT, and Digit.
| ā | Yale | ORL | LSVT | Digit | ā | Acc | Time | Acc | Time | Acc | Time | Acc | Time |
| SVM(best) | 69.3231 | 0.2981 | 80.0002 | 1.6798 | 76.1905 | 0.0142 | 95.9302 | 0.9195 | LR-SVM(best) | 79.8667 | 0.0109 | 87.0015 | 0.1575 | 77.6637 | 0.0043 | 96.3304 | 0.0972 | UMKL(+) [16] | 87.5114 | 2.5981 | 93.5798 | 18.2352 | 78.5714 | 0.0229 | 97.3571 | 4.7412 | LR-UMKL(+) | 95.8936 | 0.3018 | 83.4813 | 1.0817 | 73.8095 | 0.0152 | 98.1595 | 2.0278 | UMKL() [16] | 58.3248 | 2.7244 | 77.5022 | 20.5507 | 66.6935 | 0.0281 | 96.0904 | 7.1661 | LR-UMKL() | 93.5739 | 0.2636 | 93.4063 | 2.1836 | 67.0033 | 0.0176 | 98.4618 | 3.6392 | AMKL [17] | 85.7753 | 3.7236 | 93.8741 | 4.6854 | 80.9524 | 0.0452 | 97.4725 | 11.2138 | LR-AMKL | 94.3799 | 0.3869 | 96.9417 | 0.4592 | 88.0952 | 0.0085 | 98.6952 | 6.8804 | GMKL [18] | 86.2989 | 4.5330 | 96.2057 | 5.0070 | 85.7143 | 0.0565 | 99.1499 | 8.0774 | LR-GMKL | 95.8015 | 0.6253 | 98.5833 | 0.7761 | 88.9286 | 0.0183 | 99.4673 | 4.5972 | LMKL(sof) [19] | 87.9077 | 215.4055 | 97.0003 | 220.3122 | 85.0090 | 5.1989 | 99.8898 | 166.7978 | LR-LMKL(sof) | 97.9352 | 22.7204 | 98.9724 | 17.3791 | 86.7429 | 1.1933 | 98.3591 | 98.1812 | LMKL(sig) [19] | 88.0145 | 106.7552 | 97.0108 | 107.0911 | 88.7541 | 0.7238 | 99.3750 | 48.5914 | LR-LMKL(sig) | 98.0667 | 15.9711 | 97.9979 | 11.7240 | 92.6627 | 0.4594 | 99.5625 | 24.5927 | HMKL [20] | 63.6037 | 92.4410 | 93.5109 | 118.2340 | 80.5998 | 0.0915 | 97.6258 | 10.3559 | LR-HMKL | 91.9611 | 8.5972 | 98.6893 | 10.9368 | 85.1625 | 0.0352 | 98.3479 | 6.3959 | CMKL [21] | 86.4166 | 95.0618 | 96.0308 | 107.6940 | 79.9503 | 0.0874 | 96.5014 | 10.6074 | LR-CMKL | 94.0083 | 10.6380 | 98.4799 | 12.9746 | 93.9024 | 0.0347 | 98.9113 | 6.2696 | PMKL() [22] | 89.0035 | 6.1842 | 98.4901 | 6.9065 | 92.8571 | 0.1079 | 99.5881 | 24.8702 | LR-PMKL() | 99.1429 | 0.9053 | 99.5712 | 0.9828 | 95.9386 | 0.0573 | 100.0000 | 12.9831 | PMKL() [22] | 89.0261 | 5.3893 | 98.7533 | 6.5450 | 92.4662 | 0.1295 | 99.5046 | 21.1679 | LR-PMKL() | 98.8968 | 0.8494 | 99.5828 | 0.8911 | 95.5145 | 0.0651 | 99.7941 | 13.5108 | ANMKL() [23, 24] | 86.7210 | 6.4856 | 98.4396 | 20.4564 | 91.9827 | 0.1167 | 98.0007 | 10.3979 | LR-ANMKL() | 96.8667 | 0.7041 | 99.4643 | 2.8519 | 92.2247 | 0.0247 | 99.2850 | 6.9070 | ANMKL() [23, 24] | 86.6998 | 7.0664 | 98.2204 | 21.1615 | 93.0035 | 0.1194 | 98.0039 | 9.7753 | LR-ANMKL() | 97.2917 | 0.8863 | 99.2857 | 3.0597 | 92.5391 | 0.0224 | 99.2497 | 5.0374 |
|
|