All Tables

Table 1: Mean absolute errors on the evaluation set of 908 spectra in the SS model for different numbers of PCs retained in the reconstruction. (As PCA is done separately on the blue and red regions, the total number of inputs is twice the number of PCs.).
Table 2: Mean absolute errors on the evaluation set of 19 000 spectra in the RR model (plotted in Fig. 5). The first line is for the full data set (training and evaluation data). The second and third are just for the evaluation sets. The third line is for a model which included the four photometric colours as additional model inputs (predictors).
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 ${\it SNR}=35/1$ ).
Table 4: Globular/Open Clusters, literature values. The selection constraints applied for identification of likely members are labeled with *.
Table 5: RR: partial results. We list the mean $\mu$ and the corresponding standard deviation $\sigma$ of the difference Committee-SSPP for each of the different stellar types and parameter ranges.
Table 6: RR: partial results. We list the mean $\mu$ and the corresponding standard deviation $\sigma$ of the difference Committee-SSPP for each of the different stellar temperatures and metallicity ranges.
Table 7: SR: partial results. We list the mean $\mu$ and the corresponding standard deviation $\sigma$ of the difference Committee-SSPP for each of the different stellar types and parameter ranges.
Table 8: SR: Partial results. We list the mean $\mu$ and the corresponding standard deviation $\sigma$ of the difference Committee-SSPP for each of the different stellar temperatures and metallicity ranges.