Machine Learning on Difference Image Analysis: A comparison of methods for transient detection. (arXiv:1812.10518v1 [astro-ph.IM])
<a href="http://arxiv.org/find/astro-ph/1/au:+Sanchez_B/0/1/0/all/0/1">B. Sánchez</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+R%2E_M/0/1/0/all/0/1">M. J. Domínguez R.</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Lares_M/0/1/0/all/0/1">M. Lares</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Beroiz_M/0/1/0/all/0/1">M. Beroiz</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Cabral_J/0/1/0/all/0/1">J. B. Cabral</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Gurovich_S/0/1/0/all/0/1">S. Gurovich</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Quinones_C/0/1/0/all/0/1">C. Quiñones</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Artola_R/0/1/0/all/0/1">R. Artola</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Colazo_C/0/1/0/all/0/1">C. Colazo</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Schneiter_M/0/1/0/all/0/1">M. Schneiter</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Girardini_C/0/1/0/all/0/1">C. Girardini</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Tornatore_M/0/1/0/all/0/1">M. Tornatore</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Castellon_J/0/1/0/all/0/1">J. L. Nilo Castellón</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Lambas_D/0/1/0/all/0/1">D. García Lambas</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Diaz_M/0/1/0/all/0/1">M. C. Díaz</a>
We present a comparison of several Difference Image Analysis (DIA)
techniques, in combination with Machine Learning (ML) algorithms, applied to
the identification of optical transients associated with gravitational wave
events. Each technique is assessed based on the scoring metrics of Precision,
Recall, and their harmonic mean F1, measured on the DIA results as standalone
techniques, and also in the results after the application of ML algorithms, on
transient source injections over simulated and real data. This simulations
cover a wide range of instrumental configurations, as well as a variety of
scenarios of observation conditions, by exploring a multi dimensional set of
relevant parameters, allowing us to extract general conclusions related to the
identification of transient astrophysical events. The newest subtraction
techniques, and particularly the methodology published in Zackay et al. (2016)
are implemented in an Open Source Python package, named properimage, suitable
for many other astronomical image analyses. This together with the ML libraries
we describe, provides an effective transient detection software pipeline. Here
we study the effects of the different ML techniques, and the relative feature
importances for classification of transient candidates, and propose an optimal
combined strategy. This constitutes the basic elements of pipelines that could
be applied in searches of electromagnetic counterparts to GW sources.
We present a comparison of several Difference Image Analysis (DIA)
techniques, in combination with Machine Learning (ML) algorithms, applied to
the identification of optical transients associated with gravitational wave
events. Each technique is assessed based on the scoring metrics of Precision,
Recall, and their harmonic mean F1, measured on the DIA results as standalone
techniques, and also in the results after the application of ML algorithms, on
transient source injections over simulated and real data. This simulations
cover a wide range of instrumental configurations, as well as a variety of
scenarios of observation conditions, by exploring a multi dimensional set of
relevant parameters, allowing us to extract general conclusions related to the
identification of transient astrophysical events. The newest subtraction
techniques, and particularly the methodology published in Zackay et al. (2016)
are implemented in an Open Source Python package, named properimage, suitable
for many other astronomical image analyses. This together with the ML libraries
we describe, provides an effective transient detection software pipeline. Here
we study the effects of the different ML techniques, and the relative feature
importances for classification of transient candidates, and propose an optimal
combined strategy. This constitutes the basic elements of pipelines that could
be applied in searches of electromagnetic counterparts to GW sources.
http://arxiv.org/icons/sfx.gif
