Scalable end-to-end recurrent neural network for variable star classification

Simple item page
dc.contributor.authorBecker Troncoso, Ignacio
dc.contributor.authorPichara Baksai, Karim Elías
dc.contributor.authorCatelan, Márcio
dc.contributor.authorProtopapas, P.
dc.contributor.authorAguirre Orellana, Carlos Alfonso
dc.contributor.authorNikzat, Fatemeh
dc.date.accessioned2021-09-29T15:59:48Z
dc.date.available2021-09-29T15:59:48Z
dc.date.issued2020
dc.description.abstractDuring the last decade, considerable effort has been made to perform automatic classification of variable stars using machine-learning techniques. Traditionally, light curves are represented as a vector of descriptors or features used as input for many algorithms. Some features are computationally expensive, cannot be updated quickly and hence for large data sets such as the LSST cannot be applied. Previous work has been done to develop alternative unsupervised feature extraction algorithms for light curves, but the cost of doing so still remains high. In this work, we propose an end-to-end algorithm that automatically learns the representation of light curves that allows an accurate automatic classification. We study a series of deep learning architectures based on recurrent neural networks and test them in automated classification scenarios. Our method uses minimal data pre-processing, can be updated with a low computational cost for new observations and light curves, and can scale up to massive data sets. We transform each light curve into an input matrix representation whose elements are the differences in time and magnitude, and the outputs are classification probabilities. We test our method in three surveys: OGLE-III, Gaia, and WISE. We obtain accuracies of about 95 per cent in the main classes and 75 per cent in the majority of subclasses. We compare our results with the Random Forest classifier and obtain competitive accuracies while being faster and scalable. The analysis shows that the computational complexity of our approach grows up linearly with the light-curve size, while the traditional approach cost grows as Nlog (N).
dc.format.extent15 páginas
dc.fuente.origenOUP
dc.identifier.doi10.1093/mnras/staa350
dc.identifier.eissn1365-2966
dc.identifier.issn0035-8711
dc.identifier.urihttps://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=9111267
dc.identifier.urihttps://doi.org/10.1093/mnras/staa350
dc.identifier.wosidWOS:000525996700103
dc.information.autorucEscuela de ingeniería ; Becker Troncoso, Ignacio ; S/I ; 170384
dc.information.autorucEscuela de ingeniería ; Pichara Baksai, Karim Elías ; S/I ; 6541
dc.information.autorucInstituto de astrofísica ; Catelan, Márcio ; S/I ; 1001514
dc.information.autorucEscuela de ingeniería ; Aguirre Orellana, Carlos Alfonso ; S/I ; 213956
dc.information.autorucInstituto de física ; Nikzat, Fatemeh ; S/I ; 237450
dc.issue.numero1
dc.language.isoen
dc.nota.accesoContenido parcial
dc.pagina.final2995
dc.pagina.inicio2981
dc.publisherOUP
dc.revistaMON NOT R ASTRON SOCes_ES
dc.revistaMonthly Notices of the Royal Astronomical Society
dc.rightsacceso restringido
dc.subjectMethods: data analysis
dc.subjectAstronomical data bases: miscellaneous
dc.subjectSoftware: development
dc.subjectStars variables: general
dc.subject.ods11 Sustainable Cities and Communities
dc.subject.odspa11 Ciudades y comunidades sostenibles
dc.titleScalable end-to-end recurrent neural network for variable star classificationes_ES
dc.typeartículo
dc.volumen493
sipa.codpersvinculados170384
sipa.codpersvinculados6541
sipa.codpersvinculados1001514
sipa.codpersvinculados213956
sipa.codpersvinculados237450
Files
Collections
Artículos de revistas