J/MNRAS/428/220 Gamma-ray AGN type determination (Hassan+, 2013)
Gamma-ray active galactic nucleus type through machine-learning algorithms.
Hassan T., Mirabal N., Contreras J.L., Oya I.
<Mon. Not. R. Astron. Soc., 428, 220-225 (2013)>
=2013MNRAS.428..220H 2013MNRAS.428..220H (SIMBAD/NED BibCode)
ADC_Keywords: Gamma rays ; Active gal. nuclei
Keywords: galaxies: active
Abstract:
The Fermi Gamma-ray Space Telescope (Fermi) is producing the most
detailed inventory of the gamma-ray sky to date. Despite tremendous
achievements approximately 25 per cent of all Fermi extragalactic
sources in the Second Fermi Large Area Telescope Catalogue (2FGL) are
listed as active galactic nuclei (AGN) of uncertain type. Typically,
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 spectral properties. After training and testing on
identified/associated AGN from the 2FGL we find that 235 out of 269
AGN of uncertain type have properties compatible with gamma-ray BL
Lacertae and flat-spectrum radio quasars with accuracy rates of 85 per
cent. Additionally, direct comparison of our results with class
predictions made after following the infrared colour-colour space of
Massaro et al. shows that the agreement rate is over four-fifths for
54 overlapping sources, providing independent cross-validation. These
results can help tailor follow-up spectroscopic programmes and inform
future pointed surveys with ground-based Cherenkov telescopes.
Description:
In this paper, we employ Support Vector Machines (SVMs) and Random
Forest (RF) that embody two of the most robust supervised learning
algorithms available today.
We are interested in building classifiers that can distinguish between
two AGN classes: BL Lacs and FSRQs. In the 2FGL, there is a total set
of 1074 identified/associated AGN objects with the following labels:
'bzb' (BL Lacs), 'bzq' (FSRQs), 'agn' (other non-blazar AGN) and 'agu'
(active galaxies of uncertain type). From this global set, we group
the identified/associated blazars ('bzb' and 'bzq' labels) as
the training/testing set of our algorithms.
File Summary:
--------------------------------------------------------------------------------
FileName Lrecl Records Explanations
--------------------------------------------------------------------------------
ReadMe 80 . This file
table1.dat 32 257 *Predictions for Fermi AGN of uncertain type in
the 2FGL, ordered by RA
table3.dat 32 216 Predictions for unassociated Fermi objects
tagged as AGN by Mirabal et al. (2012, Cat.
J/MNRAS/424/L64), ordered by RA
--------------------------------------------------------------------------------
Note on table1.dat: Threshold values Pbzb<0.2 (in the case of FSRQs) and
Pbzb>0.8 (in the case of BL Lacs) must be met in both methods.
--------------------------------------------------------------------------------
See also:
J/ApJ/743/171 : The 2LAC catalog (Ackermann+, 2011)
J/ApJS/199/31 : Fermi LAT second source catalog (2FGL) (Nolan+, 2012)
J/MNRAS/424/L64 : AGN/pulsar distinction for 2FGL sources (Mirabal+, 2012)
Byte-by-byte Description of file: table1.dat table3.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 4 A4 --- --- [2FGL]
6- 18 A13 --- 2FGL 2FGL name, JHHMM.m+DDMMa
20- 23 F4.2 --- P(bzb)RF Probabibility of bzb type from Random Forest
(RF) method
25- 28 F4.2 --- P(bzb)SVM Probabibility of bzb type from Support Vector
Machines (SVMs) method
30- 32 A3 --- Type [bzq] Predicted type (1)
--------------------------------------------------------------------------------
Note (1): Types as follows:
bzb = BL Lacs
bzq = FSRQs (Flat-Spectrum Radio Quasar)
--------------------------------------------------------------------------------
History:
Copied at http://www.gae.ucm.es/?thassan/agus.html
(End) Patricia Vannier [CDS] 02-Dec-2013