Table 3: Mean absolute discrepancies (between our SR model and SSPP) calculated on the evaluation set of 38 731 real spectra (see also Fig. 7). Our models use PCA pre-processing for estimating $\log~T_{\rm eff}$ and ${\rm [Fe/H]}$ and WRS pre-processing for estimating $\log~g$ ; for the latter, PCA results are shown for comparison. Separate models were applied for low and high SNR spectra (the transition being at ${\it SNR}=35/1$ ).
Method SNR $E_{\log ~T_{\rm eff}}$ $E_{\log ~g}$ $E_{\rm [Fe/H]}$

PCA (25+25) 0.0138 0.4288 0.2606

low 0.0143 0.7549 0.3023

high 0.0136 0.3465 0.2384

WRS - 0.4459 -

low - 0.4495 -

high - 0.4450 -

**Table 3:** Mean absolute discrepancies (between our SR model and SSPP) calculated on the evaluation set of 38 731 real spectra (see also Fig. 7). Our models use PCA pre-processing for estimating $\log~T_{\rm eff}$ and ${\rm [Fe/H]}$ and WRS pre-processing for estimating $\log~g$ ; for the latter, PCA results are shown for comparison. Separate models were applied for low and high *SNR* spectra (the transition being at ${\it SNR}=35/1$ ).
Method	SNR	$E_{\log ~T_{\rm eff}}$	$E_{\log ~g}$	$E_{\rm [Fe/H]}$
PCA (25+25)		0.0138	0.4288	0.2606
	low	0.0143	0.7549	0.3023
	high	0.0136	0.3465	0.2384
WRS		-	0.4459	-
	low	-	0.4495	-
	high	-	0.4450	-

Source LaTeX | All tables | In the text