RAVE: Automatic spectral classification

The Radial Velocity Experiment (RAVE) is an ambitious all-sky spectroscopic survey to measure radial velocities and stellar atmosphere parameters (temperature, metallicity, surface gravity) of up to one million stars using the 6dF multi-object spectrograph on the 1.2-m UK Schmidt Telescope of the Anglo-Australian Observatory (AAO). We train self-organizing maps, using the continuum normalized spectra of 60 000 RAVE stars. Each node of the self-organizing map then corresponds to a template spectrum of a representative type of star in the RAVE data. This can be used in three ways, namely: [1] the identification of outliers corresponding to stars of unusual spectral type, [2] the testing and corroboration of the pipeline by the identification of artefacts and systematic errors, and [3] the construction of a natural classification scheme for the spectral types of stars. We provide illustrations of fully trained maps, together with examples of outliers corresponding both to unusual objects and to experimental artefacts. An extension of these methods incorporating kinematic data may provide a natural way of discriminating between stars belonging to different components, such as the thick and thin disks