We explore the application of artificial neural networks (ANNs) for the estimation of atmospheric parameters (T_eff_, log(g), and [Fe/H]) for Galactic F- and G-type stars. The ANNs are fed with medium-resolution ({Delta}{lambda}~1-2{AA}) nonflux-calibrated spectroscopic observations. From a sample of 279 stars with previous high-resolution determinations of metallicity and a set of (external) estimates of temperature and surface gravity, our ANNs are able to predict T_eff_ with an accuracy of {sigma}(T_eff_)=135-150K over the range 4250K<=T_eff_<=6500K, logg with an accuracy of {sigma}(logg)=0.25-0.30dex over the range 1.0<=logg<=5.0, and [Fe/H] with an accuracy {sigma}([Fe/H])=0.15-0.20dex over the range -4.0<=[Fe/H]<=0.3. Such accuracies are competitive with the results obtained by fine analysis of high-resolution spectra. It is noteworthy that the ANNs are able to obtain these results without consideration of photometric information for these stars. We have also explored the impact of the signal-to-noise ratio (S/N) on the behavior of ANNs and conclude that, when analyzed with ANNs trained on spectra of commensurate S/N, it is possible to extract physical parameter estimates of similar accuracy with stellar spectra having S/N as low as 13. Taken together, these results indicate that the ANN approach should be of primary importance for use in present and future large-scale spectroscopic surveys. The stars that comprise our study are a subset of the calibration stars used in the Beers et al. (1999, Cat. ) medium-resolution surveys.

Cone search capability for table J/ApJ/562/528/table2 (Catalog and ANN parameters for the training sample)

Cone search capability for table J/ApJ/562/528/table3 (Catalog and ANN parameters for the testing sample)