Quantitative Detection of Quartz Sandstone SiO₂ Grade Using Polarized Infrared Absorption Spectroscopy with Convolutional Neural Network Model

Table 6

Assessment of the CNN prediction results.

Input variable	R^2	RMSE	RPIQ
Original data	0.6811	0.045	2.61
First-order derivative	0.7404	0.036	3.26
Reciprocal logarithm	0.7025	0.045	2.61
MSC	0.7416	0.036	3.26
Original data-PCA	0.7742	0.033	3.56
First-order derivative-PCA	0.8143	0.031	3.79
Reciprocal logarithm-PCA	0.9065	0.023	5.11
MSC-PCA	0.8450	0.028	4.20
Original data-SPA	0.8081	0.031	3.79
First-order derivative-SPA	0.7035	0.038	3.09
Reciprocal logarithm-SPA	0.8100	0.033	3.56
MSC-SPA	0.8156	0.030	3.92