Identifying transient and variable sources in radio images

A. Rowlinson (1) , A. Stewart (2) , J.W. Broderick (3) , J.D. Swinbank (4) , A. Wijers (1) , D. Carbone (5) , Y. Cendes (6) , R. Fender (7) , A. van Der Horst (8) , G. Molenaar (9) , B. Scheers (10) , T. Staley , S. Farrell , Jean-Mathias Griessmeier (11) , M. Bell (12) , J. Eislöffel (13) , J. Law , J. van Leeuwen (14) , P. Zarka (15)

1 Anton Pannekoek Institute for Astronomy
2 LASP - Laboratory for Atmospheric and Space Physics [Boulder]
3 SIfA - Sydney Institute for Astronomy
4 University of Washington [Seattle]
5 Department of Physics and Astronomy
6 Dunlap Institute for Astronomy and Astrophysics [Toronto]
7 Oxford Astrophysics
8 Department of Physics the George Washington University
9 Department of Physics and Electronics
10 CWI - Centrum Wiskunde & Informatica
11 LPC2E - Laboratoire de Physique et Chimie de l'Environnement et de l'Espace
12 UTS - University of Technology Sydney
13 TLS - Thüringer Landessternwarte Tautenburg
14 Department of Astronomy [Berkeley]
15 LESIA (UMR_8109) - Laboratoire d'études spatiales et d'instrumentation en astrophysique

With the arrival of a number of wide-field snapshot image-plane radio transient surveys, there will be a huge influx of images in the coming years making it impossible to manually analyse the datasets. Automated pipelines to process the information stored in the images are being developed, such as the LOFAR Transients Pipeline, outputting light curves and various transient parameters. These pipelines have a number of tuneable parameters that require training to meet the survey requirements. This paper utilises both observed and simulated datasets to demonstrate different machine learning strategies that can be used to train these parameters. We use a simple anomaly detection algorithm and a penalised logistic regression algorithm. The datasets used are from LOFAR observations and we process the data using the LOFAR Transients Pipeline; however the strategies developed are applicable to any light curve datasets at different frequencies and can be adapted to different automated pipelines. These machine learning strategies are publicly available as Python tools that can be downloaded and adapted to different datasets (https://github.com/AntoniaR/TraP_ML_tools).

Keywords: Methods Data analysis Methods Statistical Radio continuum General

Domains: Sciences of the Universe [physics]; Physics [physics]; Physics [physics]; Astrophysics [astro-ph]

A. Rowlinson, A. Stewart, J.W. Broderick, J.D. Swinbank, A. Wijers, et al.. Identifying transient and variable sources in radio images. Astronomy and Computing, 2019, 27, pp.111-129. ⟨10.1016/j.ascom.2019.03.003⟩. ⟨insu-02123783⟩