J/MNRAS/490/4770    Classifying Fermi-LAT gamma-ray sources   (Kovacevic+, 2019)

Optimizing neural network techniques in classifying Fermi-LAT gamma-ray sources. Kovacevic M., Chiaro G., Cutini S., Tosti G. <Mon. Not. R. Astron. Soc., 490, 4770-4777 (2019)> =2019MNRAS.490.4770K 2019MNRAS.490.4770K (SIMBAD/NED BibCode)
ADC_Keywords: Active gal. nuclei ; BL Lac objects ; Gamma rays ; X-ray sources ; Radio sources Keywords: methods: statistical - galaxies: active - BL Lacertae objects: general - gamma-rays: galaxies Abstract: Machine learning is an automatic technique that is revolutionizing scientific research, with innovative applications and wide use in astrophysics. The aim of this study was to develop an optimized version of an Artificial Neural Network machine learning method for classifying blazar candidates of uncertain type detected by the Fermi Large Area Telescope γ-ray instrument. The final result of this study increased the classification performance by about 80 per cent with respect to previous method, leaving only 15 unclassified blazars out of 573 blazar candidates of uncertain type listed in the LAT 4-year Source Catalog. Description: In this study, we explored the possibilities to increase the performance of a neural network method previously used for the classification of uncertain blazars in Chiaro et al. (2016MNRAS.462.3180C 2016MNRAS.462.3180C, Cat. J/MNRAS/462/3180). We developed an optimized version of the original algorithm improving the selecting performance of about 80 per cent. The final result of this study left 15 uncertain blazar sources instead of 77 in Chiaro et al. (2016MNRAS.462.3180C 2016MNRAS.462.3180C, Cat. J/MNRAS/462/3180). Looking beyond γ-ray features of blazars, interesting information can be obtained from a multiwavelength study of the sources and particularly from X-ray and radio flux. In this study we tested the possibility to use those two parameters to improve the performance of the network. We did not consider any optical spectroscopy data because when considering uncertain sources, optical spectra are very often not available or not sufficiently descriptive of the nature of the source. The γ-ray flux was obtained by adding five time-integrated fluxes in five bands (0.1-0.3, 0.3-1, 1-3, 3-10, 10-100 GeV) from the 3FGL catalogue (Acero et al. 2015ApJS..218...23A 2015ApJS..218...23A, Cat. J/ApJS/218/23). Radio and X-ray data were obtained from the Fermi-LAT 4-year AGN Catalog 3LAC (Ackermann et al. 2015ApJ...810...14A 2015ApJ...810...14A, Cat. J/ApJ/810/14). Radio fluxes used were measured at frequencies of 1.4 and 0.8GHz; the X-ray fluxes were measured in the 0.1-2.4keV range. The complete list of 567 classified BCUs is presented in Table 1 in which sources are sorted by increasing likelihood of a source being a BL Lac. File Summary: -------------------------------------------------------------------------------- FileName Lrecl Records Explanations -------------------------------------------------------------------------------- ReadMe 80 . This file table1.dat 72 567 Properties of 567 BCUs used for classification in this work -------------------------------------------------------------------------------- See also: J/MNRAS/462/3180 : 3FGL Blazar of Unknown Type classification (Chiaro+, 2016) J/ApJS/218/23 : Fermi LAT third source catalog (3FGL) (Acero+, 2015) J/ApJ/810/14 : Third catalog of LAT-detected AGNs (3LAC) (Ackermann+, 2015) Byte-by-byte Description of file: table1.dat -------------------------------------------------------------------------------- Bytes Format Units Label Explanations -------------------------------------------------------------------------------- 1- 18 A18 --- Name Blazar name (3FGL JHHMM.m+DDMMe) 20- 26 F7.3 deg GLAT Galactic latitude 28- 34 F7.3 deg GLON Galactic longitude 36- 40 F5.3 --- lBLLac BL Lac likelihood from this work 42- 46 F5.3 --- PBLLac ? BL Lac precision from this work 48- 52 F5.3 --- PFSRQ ? FSRQ precision from this work 54- 59 A6 --- Class BCU classification from this work (1) 61- 65 F5.3 --- lBLLaclit BL Lac likelihood according to Chiaro et al. (2016MNRAS.462.3180C 2016MNRAS.462.3180C, Cat. J/MNRAS/462/3180) 67- 72 A6 --- Classlit BCU classification according to Chiaro et al. (2016MNRAS.462.3180C 2016MNRAS.462.3180C, Cat. J/MNRAS/462/3180) (2) -------------------------------------------------------------------------------- Note (1): Classification as follows: BCU = Blazar candidates of uncertain type (15/567) BL Lac = BL Lacertae (378/567) FSRQ = Flat Spectrum Radio Quasar (174/567) Note (2): Classification as follows: BCU = Blazar candidates of uncertain type (75/567) BL Lac = BL Lacertae (341/567) FSRQ = Flat Spectrum Radio Quasar (151/567) -------------------------------------------------------------------------------- History: From electronic version of the journal
(End) Ana Fiallos [CDS] 03-Feb-2023
The document above follows the rules of the Standard Description for Astronomical Catalogues; from this documentation it is possible to generate f77 program to load files into arrays or line by line