Automatic Classification of Time-Variable X-Ray Sources

Lo, Kitty; Farrell, Sean; Murphy, Tara; Gaensler, B. M.

doi:10.1088/0004-637x/786/1/20

Cited by 25 publications

(41 citation statements)

References 54 publications

(51 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We added extra columns to estimate the source variability. This could be done directly from the light curves, as in Lo et al (2014), but as a first approach we computed only the logarithm of the ratio between maximum and mean X-ray fluxes, in the most variable energy bands (among soft 0.3-1 keV, medium 1-2 keV, and hard 2-10 keV bands). This ratio was estimated by the quan-tity P eakRate bandi /Rate bandi or P eakRate bandi /U L bandi if Rate bandi is zero, where U L bandi stands for the 3σ upper limit on the count rate, given in 2SXPS.…”

Section: Catalogmentioning

confidence: 99%

“…These populations can then be used to answer questions regarding the hierarchical evolution of galaxies or how the earliest supermassive black holes formed (e.g., Greene 2012). One possible automated way to find such rare objects serendipitously is to search for outliers in the parameter space of the catalog (e.g., Lo et al 2014) within the framework of X-ray source classification. X-ray source classification can also be used for population studies, for example to study the X-ray luminosity function of high-mass X-ray binaries (HMXBs; Mineo et al 2012), to perform quick diagnostics on an object's nature for individual studies, or to spot unstudied objects in unexpected environments (e.g., Lin et al 2014).…”

Section: Introductionmentioning

confidence: 99%

“…Taken individually, each tree performs poorly at classifying sources, but the forest as a whole is able to perform much better. Lo et al (2014) showed that XMM-Newton variable sources could be reasonably classified thanks to different variability features computed from XMM-Newton light curves. The classification efficiency is, however, boosted as soon as the classifier takes other data into account, in particular the multiwavelength information obtained via positional crossmatch between XMM-Newton and optical or infrared catalogs (Pineau et al 2017, Salvato et al 2018.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Probabilistic classification of X-ray sources applied to Swift-XRT and XMM-Newton catalogs

Hugo

Godet

Webb

et al. 2022

A&A

View full text Add to dashboard Cite

Context. Serendipitous X-ray surveys have proven to be an efficient way to find rare objects, for example tidal disruption events, changing-look active galactic nuclei (AGN), binary quasars, ultraluminous X-ray sources, and intermediate mass black holes. With the advent of very large X-ray surveys, an automated classification of X-ray sources becomes increasingly valuable. Aims. This work proposes a revisited naive Bayes classification of the X-ray sources in the Swift-XRT and XMM-Newton catalogs into four classes – AGN, stars, X-ray binaries (XRBs), and cataclysmic variables (CVs) – based on their spatial, spectral, and timing properties and their multiwavelength counterparts. An outlier measure is used to identify objects of other natures. The classifier is optimized to maximize the classification performance of a chosen class (here XRBs), and it is adapted to data mining purposes. Methods. We augmented the X-ray catalogs with multiwavelength data, source class, and variability properties. We then built a reference sample of about 25 000 X-ray sources of known nature. From this sample, the distribution of each property was carefully estimated and taken as reference to assign probabilities of belonging to each class. The classification was then performed on the whole catalog, combining the information from each property. Results. Using the algorithm on the Swift reference sample, we retrieved 99%, 98%, 92%, and 34% of AGN, stars, XRBs, and CVs, respectively, and the false positive rates are 3%, 1%, 9%, and 15%. Similar results are obtained on XMM sources. When applied to a carefully selected test sample, representing 55% of the X-ray catalog, the classification gives consistent results in terms of distributions of source properties. A substantial fraction of sources not belonging to any class is efficiently retrieved using the outlier measure, as well as AGN and stars with properties deviating from the bulk of their class. Our algorithm is then compared to a random forest method; the two showed similar performances, but the algorithm presented in this paper improved insight into the grounds of each classification. Conclusions. This robust classification method can be tailored to include additional or different source classes and can be applied to other X-ray catalogs. The transparency of the classification compared to other methods makes it a useful tool in the search for homogeneous populations or rare source types, including multi-messenger events. Such a tool will be increasingly valuable with the development of surveys of unprecedented size, such as LSST, SKA, and Athena, and the search for counterparts of multi-messenger events.

show abstract

Section: Catalogmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Probabilistic classification of X-ray sources applied to Swift-XRT and XMM-Newton catalogs

Hugo

Godet

Webb

et al. 2022

A&A

View full text Add to dashboard Cite

show abstract

“…Other known parameters, such as optical to X-ray flux ratio, are used afterwards to discriminate between source classes (Pineau et al 2011;Lin et al 2012). More recent works have applied machine learning algorithms where timing parameters are the major classifying feature (Lo et al 2014;Farrell et al 2015). Around 27.1% of the EXOD detected sources did not have a previously generated light curve and thus no variability classification.…”

Section: Discoveriesmentioning

confidence: 99%

The EXOD search for faint transients in XMM-Newton observations: Method and discovery of four extragalactic Type I X-ray bursters

Pastor-Marazuela

Webb

Wojtowicz

et al. 2020

A&A

View full text Add to dashboard Cite

Context. The observations carried out with XMM-Newton have produced a very extensive X-ray source catalogue in which the standard pipeline determines the variability of sufficiently bright sources through χ2 and fractional variability tests. Faint sources, however, are not automatically checked for variability, and this means that faint, short timescale transients are overlooked. From dedicated X-ray searches, as well as optical and radio archive searches, we know that some such dim sources can still be identified with high confidence. Aims. Our goal is to find new faint, fast transients in XMM-Newton EPIC-pn observations. To that end we created the EPIC-pn XMM-Newton outburst detector (EXOD) algorithm, which we run on the EPIC-pn full-frame data available in the 3XMM-DR8 catalogue. Methods. In EXOD, we computed the variability of the whole field of view by first binning in time the counts detected in each pixel of the detector. We next computed the difference between the median and maximal number of counts in each time bin and pixel to detect variability. We applied EXOD to 5751 observations in the full frame mode and compared the variability of the detected sources to the standard χ2 and Kolmogorov–Smirnov (KS) variability tests. Results. The algorithm is able to detect periodic and aperiodic variability, with both short and long flares. Of the sources detected by EXOD, 60−95% are also shown to be variable by the standard χ2 and KS tests. EXOD computes the variability over the entire field of view faster than the light curve generation takes for all the individual sources. We detect a total of 2961 X-ray variable sources. After removing the spurious detections, we obtain a net number of 2536 variable sources. Of these we investigate the nature of 35 sources with no previously confirmed classification. Amongst the new sources, we find stellar flares and AGNs, in addition to four extragalactic type I X-ray bursters that double the known neutron-star population in M 31. Conclusions. This algorithm is a powerful tool for the prompt detection of interesting variable sources in XMM-Newton observations. EXOD also detects fast transients that other variability tests would classify as non-variable due to their short duration and low number of counts. This is of increasing importance for the multi-messenger detection of transient sources. Finally, EXOD allows us to identify the nature of compact objects through their variability and to detect rare compact objects. We demonstrate this through the discovery of four extragalactic neutron-star low-mass X-ray binaries, doubling the number of known neutron stars in M 31.

show abstract

“…Gopalan et al (2015) expanded the colour-colour-intensity diagram classification technique by applying a supervised learning algorithm as a method of demarcating systems containing black holes, pulsating neutron stars, or non-pulsating neutron stars. Lo et al (2014) employed the random forest algorithm to classify time-varying X-ray sources in the Second XMM-Newton Serendipitous Source Catalog using X-ray photometric time series and spectra and multiwavelength information. Their results indicated a high classification accuracy and the ability to detect unusual objects in their test sample.…”

Section: Machine Learning For X-ray Source Classificationmentioning

confidence: 99%

Identifying new X-ray binary candidates in M31 using random forest classification

Arnason

Barmby

Vulic

2020

Monthly Notices of the Royal Astronomical Society

View full text Add to dashboard Cite

Identifying X-ray binary (XRB) candidates in nearby galaxies requires distinguishing them from possible contaminants including foreground stars and background active galactic nuclei. This work investigates the use of supervised machine learning algorithms to identify high-probability X-ray binary candidates. Using a catalogue of 943 Chandra X-ray sources in the Andromeda galaxy, we trained and tested several classification algorithms using the X-ray properties of 163 sources with previously known types. Amongst the algorithms tested, we find that random forest classifiers give the best performance and work better in a binary classification (XRB/non-XRB) context compared to the use of multiple classes. Evaluating our method by comparing with classifications from visible-light and hard X-ray observations as part of the Panchromatic Hubble Andromeda Treasury, we find compatibility at the 90% level, although we caution that the number of source in common is rather small. The estimated probability that an object is an X-ray binary agrees well between the random forest binary and multiclass approaches and we find that the classifications with the highest confidence are in the X-ray binary class. The most discriminating X-ray bands for classification are the 1.7-2.8, 0.5-1.0, 2.0-4.0, and 2.0-7.0 keV photon flux ratios. Of the 780 unclassified sources in the Andromeda catalogue, we identify 16 new high-probability X-ray binary candidates and tabulate their properties for follow-up.

show abstract

Automatic Classification of Time-Variable X-Ray Sources

Cited by 25 publications

References 54 publications

Probabilistic classification of X-ray sources applied to Swift-XRT and XMM-Newton catalogs

Probabilistic classification of X-ray sources applied to Swift-XRT and XMM-Newton catalogs

The EXOD search for faint transients in XMM-Newton observations: Method and discovery of four extragalactic Type I X-ray bursters

Identifying new X-ray binary candidates in M31 using random forest classification

Contact Info

Product

Resources

About