Soft Sensor Modeling Method Based on SPA-GWO-SVR for Marine Protease Fermentation Process

Table 4

Quality prediction results based on SPA-GWO-SVR soft sensor model.

Key variables	Variable selection method	Training set		Test set
		R^2	RMSE	R^2	RMSE
X	Full	0.9783	1.8196	0.9763	1.8254
	GRA	0.9870	1.0641	0.9841	1.0897
	SPA	0.9945	1.0368	0.9933	1.0583
S	Full	0.9707	1.0026	0.9698	1.0254
	GRA	0.9811	0.6324	0.9786	0.6580
	SPA	0.9952	0.5066	0.9940	0.5561
P	Full	0.9794	0.0811	0.9781	0.0864
	GRA	0.9810	0.0301	0.9812	0.0348
	SPA	0.9982	0.0156	0.9971	0.0164