Examining the Potential Environmental Controls of Underground CO₂ Concentration in Arid Regions by an SVD-PCA-ANN Preview Model

Table 1

Comparison of the SVD-PCA-ANN model with SAE, SVM, and LSTM.

Accuracy evaluation index			RMSE	RPD
Training	SVD-PCA-ANN	0.8874	0.3351	2.7929
	SVM	0.1209	1.0552	0.7001
	LSTM	0.3375	0.8087	0.5367
	SAE	0.0074	0.6074	0.1999
	SVD-based PCA	0.4890	1.2330	2.9422
Validation	SVD-PCA-ANN	0.8197	0.4121	1.9069
	SVM	0.4800	1.2130	0.2723
	LSTM	0.3271	0.8253	0.5366
	SAE	0.1698	0.9435	0.1289
	SVD-based PCA	0.5028	1.2468	2.9516
Testing	SVD-PCA-ANN	0.7661	0.4257	1.9043
	SVM	0.6070	1.0491	0.6071
	LSTM	0.2722	0.8096	0.5337
	SAE	0.3713	0.9435	0.2439
	SVD-based PCA	0.4931	1.2328	2.4960