TABLE 2.
Comparisons of SSC-LRR with other methods for cancer classification on the R-GCM and MBD datasets *
| Datasets | Method | REC | FPR | MCC | Overall Acc (Q) | General corr (GC) |
|---|---|---|---|---|---|---|
| R-GCM | SVM | 0.510 (1.79×10−7) | 0.052 (1.25×10−4) | 0.503 (5.55×10−7) | 0.554 (3.37×10−5) | 0618 (1.34×10−5) |
| RPCA+SVM | 0.521 (1.20×10−4) | 0.051 (3.79×10−6) | 0.512 (3.52×10−6) | 0.562 (8.96×10−7) | 0.624 (1.48×10−6) | |
| Label Propagation | 0.324 (5.26×10−6) | 0.056 (2.85×10−4) | 0.308 (2.96×10−7) | 0.409 (2.34×10−4) | 0.445 (2.79×10−4) | |
| Semi-PNMF | 0.658 (8.07×10−4) | 0.029 (8.64×10−5) | 0.633 (9.03×10−5) | 0.707 (3.17×10−5) | 0.710 (5.10×10−4) | |
| SSC-LRR | 0.715 | 0.023 | 0.762 | 0.753 | 0.762 | |
| MBD | SVM | 0.636 (1.56×10−5) | 0.167 (3.70×10−3) | 0.504 (1.34×10−3) | 0.471 (1.34×10−4) | 0.651 (7.67×10−4) |
| RPCA+SVM | 0.451 (1.16×10−8) | 0.175 (6.92×10−6) | 0.279 (5.85×10−8) | 0.653 (6.35×10−8) | 0.385 (2.28×10−8) | |
| Label Propagation | 0.553 (9.99×10−5) | 0.184 (8.94×10−4) | 0.348 (3.77×10−5) | 0.646 (7.85×10−3) | 0.415 (1.09×10−4) | |
| Semi-PNMF | 0.884 (4.62×10−2) | 0.106 (3.96×10−2) | 0.635 (9.78×10−3) | 0.708 (3.49×10−2) | 0.676 (1.58×10−3) | |
| SSC-LRR | 0.959 | 0.033 | 0.888 | 0.897 | 0.920 |
Parameter M for splitting benchmark dataset is set to 6. Values in parentheses are p-values in student t-test relative to SSC-LRR.