Machine-learning classifiers for Fermi AGN

Abstract

The Fermi Gamma-ray Space Telescope is generating the most detailed map of the gamma-ray sky. While tremendously successful, approximately 25% of all associated Fermi extragalactic sources in the Second Fermi LAT Catalog (2FGL) are listed as active galactic nuclei (AGN) of uncertain type. Most of these are suspected blazar candidates without a conclusive optical spectrum or lacking spectroscopic observations. Here, we explore the use of machine-learning algorithms – Random Forests and Support Vector Machines – to predict specific AGN subclass based on observed gamma-ray properties.