Published online by Cambridge University Press: 15 February 2011
Classification and parametrization of unresolved galaxies to be detected by Gaia is thegoal of our contribution to DPAC/CU8. Gaia will observe more than several times106 unresolved galaxies seen almost as point sources. For the classificationand parametrization of these sources we have undertaken the following tasks: a) providinglibraries of synthetic galaxy spectra (produced under the PÉGASE.2 code of galaxyevolution) and a semi-empirical library using observed galaxy spectra from SDSS dataarchive. b) Developping an appropriate algorithm called Unresolved Galaxy Classifier(UGC). The algorithm is being developed implementing the Support Vector Machines (SVM), asupervised training technique. This algorithm utilizes Gaia simulated BP/RP galaxy spectraof our synthetic galaxy library.