Analysis of medium resolution spectra by automated methods - application to M55 and omega Centauri

We have employed feedforward neural networks trained on synthetic spectra in the range 3800 to 5600 angstrom with resolutions of similar to 2-3 angstrom to determine metallicities from spectra of about 1000 main-sequence turn-off, subgiant and red giant stars in the globular clusters M55 and omega Cen. The overall metallicity accuracies are of the order of 0.15 to 0.2 dex. In addition, we tested how well the stellar parameters log g and T-eff can be retrieved from such data without additional colour or photometric information. We find overall uncertainties of 0.3 to 0.4 dex for log g and 140 to 190 K for T-eff. In order to obtain some measure of uncertainty for the determined values of [Fe/H], log g and T-eff, we applied the bootstrap method for the first time to neural networks for this kind of parametrization problem. The distribution of metallicities for stars in omega Cen clearly shows a large spread in agreement with the well known multiple stellar populations in this cluster.