Applying machine learning to catalogue matching in astrophysics

Abstract

We present the results of applying automated machine learning techniques to the problem of matching different object catalogues in astrophysics. In this study, we take two partially matched catalogues where one of the two catalogues has a large positional uncertainty. The two catalogues we used here were taken from the H i Parkes All Sky Survey (HIPASS) and SuperCOSMOS optical survey. Previous work had matched 44 per cent (1887 objects) of HIPASS to the SuperCOSMOS catalogue. A supervised learning algorithm was then applied to construct a model of the matched portion of our catalogue. Validation of the model shows that we achieved a good classification performance (99.12 per cent correct). Applying this model to the unmatched portion of the catalogue found 1209 new matches. This increases the catalogue size from 1887 matched objects to 3096. The combination of these procedures yields a catalogue that is 72 per cent matched.

BAYESIAN MATCHING FOR X-RAY AND INFRARED SOURCES IN THE MYStIX PROJECT

“…Such an approach fits within a broad theme in the literature of using not only the counterpart's celestial position, but some property such as its color, morphology, or spectral energy distribution (e.g., Roseboom et al 2009), or even a combination of several properties (Rohde et al 2005(Rohde et al , 2006 to assign the likelihood of it being the counterpart. Although we do not know of such techniques being used for Galactic astronomy before, they are widely used in extra-galactic work, and most techniques trace their origins back to an increasingly influential paper by Sutherland & Saunders (1992).…”

AstroDAS: Sharing Assertions Across Astronomy Catalogues Through Distributed Annotation

“…In that case, spatial proximity alone cannot judge between the potential counterparts from A of each source in B, and it is necessary to either introduce prior astrophysical knowledge to help identify the most likely match, or use a machine learning algorithm [9,10]. Machine learning algorithms deduce relationships between the properties of the sources in the two catalogues and can aid the finding of associations between them.…”

Applying the Support Vector Machine Method to Matching IRAS and SDSS Catalogues

“…Cao et al, 2006;Rohde et al, 2005Rohde et al, , 2006, especially for analyzing the vast amount of data being collected by sky surveys (e.g., SDSS). Simple matching approaches such as the 'closest match,' which depend on positions only, are considered adequate only when the positional uncertainties of the matched catalogues are all very small (e.g., matching SDSS and Spitzer IRAC catalogues, Wu et al, 2005).…”

“…Rohde et al (2005) applied automated machine learning techniques to the problem of matching catalogues where one of the two has large positional uncertainties. In their study, a model was first constructed based on a supervised learning algorithm and a set of training data.…”