A Comparison of Photometric Redshift Techniques for Large Radio Surveys. (arXiv:1902.05188v1 [astro-ph.IM])
<a href="http://arxiv.org/find/astro-ph/1/au:+Norris_R/0/1/0/all/0/1">Ray P. Norris</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Salvato_M/0/1/0/all/0/1">M. Salvato</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Longo_G/0/1/0/all/0/1">G. Longo</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Brescia_M/0/1/0/all/0/1">M. Brescia</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Budavari_T/0/1/0/all/0/1">T. Budavari</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Carliles_S/0/1/0/all/0/1">S. Carliles</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Cavuoti_S/0/1/0/all/0/1">S. Cavuoti</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Farrah_D/0/1/0/all/0/1">D. Farrah</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Geach_J/0/1/0/all/0/1">J. Geach</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Luken_K/0/1/0/all/0/1">K. Luken</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Musaeva_A/0/1/0/all/0/1">A. Musaeva</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Polsterer_K/0/1/0/all/0/1">K. Polsterer</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Riccio_G/0/1/0/all/0/1">G. Riccio</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Seymour_N/0/1/0/all/0/1">N. Seymour</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Smolcic_V/0/1/0/all/0/1">V. Smol&#x10d;i&#x107;</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Vaccari_M/0/1/0/all/0/1">M. Vaccari</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Zinn_P/0/1/0/all/0/1">P. Zinn</a>

Future radio surveys will generate catalogues of tens of millions of radio
sources, for which redshift estimates will be essential to achieve many of the
science goals. However, spectroscopic data will be available for only a small
fraction of these sources, and in most cases even the optical and infrared
photometry will be of limited quality. Furthermore, radio sources tend to be at
higher redshift than most optical sources and so a significant fraction of
radio sources hosts differ from those for which most photometric redshift
templates are designed. We therefore need to develop new techniques for
estimating the redshifts of radio sources. As a starting point in this process,
we evaluate a number of machine-learning techniques for estimating redshift,
together with a conventional template-fitting technique. We pay special
attention to how the performance is affected by the incompleteness of the
training sample and by sparseness of the parameter space or by limited
availability of ancillary multi-wavelength data. As expected, we find that the
quality of the photometric-redshift degrades as the quality of the photometry
decreases, but that even with the limited quality of photometry available for
all sky-surveys, useful redshift information is available for the majority of
sources, particularly at low redshift. We find that a template-fitting
technique performs best with high-quality and almost complete multi-band
photometry, especially if radio sources that are also X-ray emitting are
treated separately. When we reduced the quality of photometry to match that
available for the EMU all-sky radio survey, the quality of the template-fitting
degraded and became comparable to some of the machine learning methods. Machine
learning techniques currently perform better at low redshift than at high
redshift, because of incompleteness of the currently available training data at
high redshifts.

Future radio surveys will generate catalogues of tens of millions of radio
sources, for which redshift estimates will be essential to achieve many of the
science goals. However, spectroscopic data will be available for only a small
fraction of these sources, and in most cases even the optical and infrared
photometry will be of limited quality. Furthermore, radio sources tend to be at
higher redshift than most optical sources and so a significant fraction of
radio sources hosts differ from those for which most photometric redshift
templates are designed. We therefore need to develop new techniques for
estimating the redshifts of radio sources. As a starting point in this process,
we evaluate a number of machine-learning techniques for estimating redshift,
together with a conventional template-fitting technique. We pay special
attention to how the performance is affected by the incompleteness of the
training sample and by sparseness of the parameter space or by limited
availability of ancillary multi-wavelength data. As expected, we find that the
quality of the photometric-redshift degrades as the quality of the photometry
decreases, but that even with the limited quality of photometry available for
all sky-surveys, useful redshift information is available for the majority of
sources, particularly at low redshift. We find that a template-fitting
technique performs best with high-quality and almost complete multi-band
photometry, especially if radio sources that are also X-ray emitting are
treated separately. When we reduced the quality of photometry to match that
available for the EMU all-sky radio survey, the quality of the template-fitting
degraded and became comparable to some of the machine learning methods. Machine
learning techniques currently perform better at low redshift than at high
redshift, because of incompleteness of the currently available training data at
high redshifts.

http://arxiv.org/icons/sfx.gif