S. Bressan | Searcharxiv

arXiv2012

Herschel-ATLAS: towards a sample of ~1000 strongly-lensed galaxies

While the selection of strongly lensed galaxies with 500μm flux density S_500>100 mJy has proven to be rather straightforward (Negrello et al. 2010), for many applications it is important to analyze samples larger than the ones obtained when confining ourselves to such a bright limit. Moreover, only by probing to fainter flux densities is possible to exploit strong lensing to investigate the bulk of the high-z star-forming galaxy population. We describe HALOS (the Herschel-ATLAS Lensed Objects Selection), a method for efficiently selecting fainter candidate strongly lensed galaxies, reaching a surface density of ~1.5-2 deg^-2, i.e. a factor of about 4 to 6 higher than that at the 100 mJy flux limit. HALOS will allow the selection of up to ~1000 candidate strongly lensed galaxies (with amplifications μ>2) over the full H-ATLAS survey area. Applying HALOS to the H-ATLAS Science Demonstration Phase field (~14.4 deg^2) we find 31 candidate strongly lensed galaxies, whose candidate lenses are identified in the VIKING near-infrared catalog. Using the available information on candidate sources and candidate lenses we tentatively estimate a ~72% purity of the sample. The redshift distribution of the candidate lensed sources is close to that reported for most previous surveys for lensed galaxies, while that of candidate lenses extends to substantially higher redshifts than found in the other surveys. The counts of candidate strongly lensed galaxies are also in good agreement with model predictions (Lapi et al. 2011). Even though a key ingredient of the method is the deep near-infrared VIKING photometry, we show that H-ATLAS data alone allow the selection of a similarly deep sample of candidate strongly lensed galaxies with an efficiency close to 50%; a slightly lower surface density (~1.45 deg^-2) can be reached with a ~70% efficiency.

astro-ph.CO↗

arXiv2004

Computer science approach to the stellar fabric of violent starforming regions in AGN

In order to analyse the large numbers of Seyfert galaxy spectra available at present, we are testing new techniques to derive their physical parameters fastly and accurately. We present an experiment on such a new technique to segregate old and young stellar populations in galactic spectra using machine learning methods. We used an ensemble of classifiers, each classifier in the ensemble specializes in young or old populations and was trained with locally weighted regression and tested using ten-fold cross-validation. Since the relevant information concentrates in certain regions of the spectra we used the method of sequential floating backward selection offline for feature selection. Very interestingly, the application to Seyfert galaxies proved that this technique is very insensitive to the dilution by the Active Galactic Nucleus (AGN) continuum. Comparing with exhaustive search we concluded that both methods are similar in terms of accuracy but the machine learning method is faster by about two orders of magnitude.

astro-ph↗