Towards automatic classification of all WISE sources

Abstract

Context. The Wide-field Infrared Survey Explorer (WISE) has detected hundreds of millions of sources over the entire sky. Classifying them reliably is, however, a challenging task owing to degeneracies in WISE multicolour space and low levels of detection in its two longest-wavelength bandpasses. Simple colour cuts are often not sufficient; for satisfactory levels of completeness and purity, more sophisticated classification methods are needed. Aims. Here we aim to obtain comprehensive and reliable star, galaxy, and quasar catalogues based on automatic source classification in full-sky WISE data. This means that the final classification will employ only parameters available from WISE itself, in particular those which are reliably measured for the majority of sources. Methods. For the automatic classification we applied a supervised machine learning algorithm, support vector machines (SVM). It requires a training sample with relevant classes already identified, and we chose to use the SDSS spectroscopic dataset (DR10) for that purpose. We tested the performance of two kernels used by the classifier, and determined the minimum number of sources in the training set required to achieve stable classification, as well as the minimum dimension of the parameter space. We also tested SVM classification accuracy as a function of extinction and apparent magnitude. Thus, the calibrated classifier was finally applied to all-sky WISE data, flux-limited to 16 mag (Vega) in the 3.4 µm channel. Results. By calibrating on the test data drawn from SDSS, we first established that a polynomial kernel is preferred over a radial one for this particular dataset. Next, using three classification parameters (W1 magnitude, W1 − W2 colour, and a differential aperture magnitude) we obtained very good classification efficiency in all the tests. At the bright end, the completeness for stars and galaxies reaches ∼95%, deteriorating to ∼80% at W1 = 16 mag, while for quasars it stays at a level of ∼95% independently of magnitude. Similar numbers are obtained for purity. Application of the classifier to full-sky WISE data and appropriate a posteriori cleaning allowed us to obtain catalogues of star and galaxy candidates that appear reliable. However, the sources flagged by the classifier as "quasars" are in fact dominated by dusty galaxies; they also exhibit contamination from sources located mainly at low ecliptic latitudes, consistent with solar system objects.

Galaxy and Mass Assembly (GAMA): Exploring the WISE Web in G12

“…unresolved point sources-except, of course, for the tiny population of resolved extragalactic sources-and may share similar color properties in certain broad bands (see, for example, Yan et al 2013;Kurcz et al 2016). Kinematic information, which may be definitive, such as reliable radial and transverse motions, is difficult and expensive to acquire.…”

“…We caution that the same cannot be said for fields closer to the Galactic Plane, where exponentially increasing numbers of stars completely overwhelm the relatively clean star-galaxy separation presented here. Photometric error scatters stars across the CMDs, notably with K and M dwarfs (e.g., Figure 6(b)), creating degeneracies that are very difficult to break without additional optical and near-infrared color phase space information-see Bilicki et al (2016) and Kurcz et al (2016) for an all-sky analysis of star-galaxy separation using optical, near-infrared, and mid-infrared colors. Below and in Section 3.5 we consider the completeness of the counts.…”

A Machine-learning Method for Identifying Multiwavelength Counterparts of Submillimeter Galaxies: Training and Testing Using AS2UDS and ALESS

“…Automatized classification using machine-learning algorithms has recently gained popularity in astronomy and has been applied to a number of problems, including star/galaxy/quasar classification (Bloom et al 2012;Solarz et al 2012;Małek et al 2013;Kurcz et al 2016) and the identification of different types of supernovae (du Buisson et al 2015;Lochner et al 2016).…”

Applications of machine-learning algorithms for infrared colour selection of Galactic Wolf–Rayet stars

“…The current training set is at least one order of magnitude smaller than what is typical in other machine-learning classification studies (e.g., Kurcz et al 2016;Marton et al 2016). With these low numbers, the use of complex machine-learning classifiers, such as Artificial Neural Networks (ANN, Jeffrey & Rosner 1986), is out of the question.…”

“…These methods use the statistical information contained in the infrared colours for a set of known objects, including non-WR stellar populations which are frequently confused as WR candidates due to similar colours, to providing an automated classification of the unknown objects. The use of supervised machinelearning methods in astronomy has rapidly increased over the last decade, e.g., for automated classification of celestial objects in large catalogues and all-sky surveys (Malek et al 2013;Kurcz et al 2016;Lochner et al 2016), photometric redshift estimation of galaxies (Tagliaferri et al 2003;Lima et al 2008;Sheldon et al 2012;Heinis et al 2016), morphological galaxy classification (Banerji et al 2010;Shamir et al 2013;Kuminski et al 2014;Pasquato & Chung 2016) and candidate type of object selection (Bailey et al 2007;Yèche et al 2010;Hsieh & Lai 2013;Marton et al 2016). To our knowledge, this is the first time that machine-learning methods are used to classify objects in this colour space defined by J, H, K s , [3.6], [4.5], [5.8] and [8.0] photometric bands.…”