Speaker: Artem Poliszczuk
Title: Fuzzy support vector machine application to data mining in sky surveys.
Abstract: The first application of the fuzzy support vector machine (FSVM) algorithm as the automated classification tool for astronomical catalogs will be presented. This new approach allows to perform a more trustable classification of astronomical sources by making use of the measurement uncertainties. The performance of different versions of the SVM algorithm is examined on the AKARI-NEP data and the resultant catalog of infrared-selected galaxies is presented.