Research Article
Robust Proximal Support Vector Regression Based on Maximum Correntropy Criterion
Table 5
Experimental results on Concrete and Abalone without outliers (I) and with outliers (II).
| | Dataset | Algorithm | () | RMSE | MAE | SSE/SST | SSR/SST |
| | Concrete (1030 × 8) (500, 530) | I | LSSVR | () | 6.2625(1) | 4.5672(2) | 0.1381(1.5) | 0.8974(4) | | PSVR | () | 6.2636(3) | 4.5684(4) | 0.1382(3.5) | 0.8975(3) | | WLSSVR-H | () | 6.3971(6) | 4.5935(6) | 0.1443(6) | 0.9178(1) | | WLSSVR-L | () | 6.3278(5) | 4.5636(1) | 0.1411(5) | 0.9024(2) | | RLSSVR-MCC | () | 6.2633(2) | 4.5681(3) | 0.1381(1.5) | 0.8972(5.5) | | RPSVR-MCC | () | 6.2644(4) | 4.5693(5) | 0.1382(3.5) | 0.8972(5.5) | | II | LSSVR | () | 7.0293(5) | 5.1977(5) | 0.1740(5.5) | 0.9005(4) | | PSVR | () | 7.0294(6) | 5.1983(6) | 0.1740(5.5) | 0.9004(5) | | WLSSVR-H | () | 7.0126(4) | 5.1833(4) | 0.1730(4) | 0.8951(6) | | WLSSVR-L | () | 6.9237(3) | 5.0999(3) | 0.1687(3) | 0.9034(3) | | RLSSVR-MCC | () | 6.9051(1) | 5.0862(1) | 0.1679(1) | 0.9159(1.5) | | RPSVR-MCC | () | 6.9060(2) | 5.0873(2) | 0.1680(2) | 0.9159(1.5) |
| | Abalone (4177 × 7) (3000, 1177) | I | LSSVR | () | 2.1066(4) | 1.5163(6) | 0.4294(4) | 0.5688(1) | | PSVR | () | 2.1057(3) | 1.5158(5) | 0.4290(3) | 0.5684(2) | | WLSSVR-H | () | 2.1240(6) | 1.4863(1) | 0.4366(6) | 0.5182(6) | | WLSSVR-L | () | 2.1086(5) | 1.4915(2) | 0.4302 (5) | 0.5459(5) | | RLSSVR-MCC | () | 2.1050(2) | 1.5068(4) | 0.4287(2) | 0.5530(3) | | RPSVR-MCC | () | 2.1040(1) | 1.5063(3) | 0.4283(1) | 0.5525(4) | | II | LSSVR | () | 2.1319(6) | 1.5454(5) | 0.4399(6) | 0.5616(1) | | PSVR | () | 2.1313(3.5) | 1.5457(6) | 0.4396(3.5) | 0.5608(3) | | WLSSVR-H | () | 2.1313(3.5) | 1.5370(2) | 0.4396(3.5) | 0.5534(5) | | WLSSVR-L | () | 2.1304(2) | 1.5354(1) | 0.4393(2) | 0.5531(6) | | RLSSVR-MCC | () | 2.1315(5) | 1.5437(4) | 0.4397(5) | 0.5576(4) | | RPSVR-MCC | () | 2.1300(1) | 1.5389(3) | 0.4391(1) | 0.5614(2) |
|
|
