Research Article
Ensemble Just-In-Time Learning-Based Soft Sensor for Mooney Viscosity Prediction in an Industrial Rubber Mixing Process
Table 2
Comparison of Mooney viscosity prediction results using different soft sensors (
= 10 for JIT learning).
| | No. | Method | RMSE | RRMSE (%) | R2 |
| | 1 | PLS | 7.3298 | 11.7026 | 0.8002 | | 2 | GPR | 4.2628 | 5.9354 | 0.9324 | | 3 | GMMGPR | 3.4606 | 4.8077 | 0.9555 | | 4 | JITGPR (ED similarity) | 3.1561 | 4.2700 | 0.9630 | | 5 | JITGPR (cosine similarity) | 3.2053 | 4.3494 | 0.9618 | | 6 | JITGPR (CWD similarity) | 3.6552 | 5.2370 | 0.9503 | | 7 | JITGPR (CC similarity) | 3.2029 | 4.3313 | 0.9618 | | 8 | SP-EJITGPR (SAR) | 3.2127 | 4.4105 | 0.9616 | | 9 | SP-EJITGPR (PLS stacking) | 3.9916 | 5.4067 | 0.9407 | | 10 | SP-EJITGPR (GPR stacking) | 3.7792 | 5.0508 | 0.9469 | | 11 | SP-EJITGPR (FMM) | 3.0670 | 4.2073 | 0.9650 | | 12 | MP-EJITGPR (SAR) | 4.3769 | 6.0332 | 0.9288 | | 13 | MP-EJITGPR (PLS stacking) | 4.1966 | 5.3943 | 0.9345 | | 14 | MP-EJITGPR (GPR stacking) | 3.7819 | 4.9735 | 0.9468 | | 15 | MP-EJITGPR (FMM) | 2.9202 | 3.9085 | 0.9683 |
|
|
ED : Euclidean distance; CWD: covariance weighted distance; CC: correlation coefficient; SAR: simple averaging rule; FMM: finite mixture mechanism; SP: similarity perturbation; and MP: multimodal perturbation.
|
