Busca avançada
Ano de início
(Referência obtida automaticamente do Web of Science, por meio da informação sobre o financiamento pela FAPESP e o número do processo correspondente, incluída na publicação pelos autores.)

Optimizing spectroscopic follow-up strategies for supernova photometric classification with active learning

Texto completo
Mostrar menos -
Ishida, E. E. O. [1] ; Beck, R. [2, 3] ; Gonzalez-Gaitan, S. [4] ; de Souza, R. S. [5] ; Krone-Martins, A. [6] ; Barrett, J. W. [7, 8] ; Kennamer, N. [9] ; Vilalta, R. [10] ; Burgess, J. M. [11] ; Quint, B. [12] ; Vitorelli, A. Z. [13] ; Mahabal, A. [14] ; Gangler, E. [1] ; Collaboration, COIN
Número total de Autores: 14
Afiliação do(s) autor(es):
Mostrar menos -
[1] Univ Clermont Auvergne, CNRS, IN2P3, LPC, F-63000 Clermont Ferrand - France
[2] Univ Hawaii, Inst Astron, 2680 Woodlawn Dr, Honolulu, HI 96822 - USA
[3] Eotvos Lorand Univ, Dept Phys Complex Syst, Pf 32, H-1518 Budapest - Hungary
[4] Univ Lisbon, Inst Super Tecn, CENTRA, COSTAR, Av Rovisco Pais 1, P-1049001 Lisbon - Portugal
[5] Univ N Carolina, Dept Phys & Astron, Chapel Hill, NC 27599 - USA
[6] Univ Lisbon, Fac Ciencias, CENTRA, SIM, Ed C8, P-1749016 Lisbon - Portugal
[7] Univ Birmingham, Sch Phys & Astron, Birmingham B15 2TT, W Midlands - England
[8] Univ Birmingham, Inst Gravitat Wave Astron, Birmingham B15 2TT, W Midlands - England
[9] Univ Calif Irvine, Dept Comp Sci, Donald Bren Hall, Irvine, CA 92617 - USA
[10] Univ Houston, Dept Comp Sci, 3551 Cullen Blvd, 501 PGH, Houston, TX 77204 - USA
[11] Max Planck Inst Extraterr Phys, Giessenbachstr, D-85748 Garching - Germany
[12] SOAR Telescope, AURA O, Colina El Pino S-N, Casila 603, La Serena - Chile
[13] Univ Sao Paulo, Inst Astron Geofis & Ciencias Atmosfer, Sao Paulo, SP - Brazil
[14] CALTECH, Ctr Data Driven Discovery, Pasadena, CA 91125 - USA
Número total de Afiliações: 14
Tipo de documento: Artigo Científico
Fonte: Monthly Notices of the Royal Astronomical Society; v. 483, n. 1, p. 2-18, FEB 2019.
Citações Web of Science: 7

We report a framework for spectroscopic follow-up design for optimizing supernova photometric classification. The strategy accounts for the unavoidable mismatch between spectroscopic and photometric samples, and can be used even in the beginning of a new survey - without any initial training set. The framework falls under the umbrella of active learning (AL), a class of algorithms that aims to minimize labelling costs by identifying a few, carefully chosen, objects that have high potential in improving the classifier predictions. As a proof of concept, we use the simulated data released after the SuperNova Photometric Classification Challenge (SNPCC) and a random forest classifier. Our results show that, using only 12 per cent the number of training objects in the SNPCC spectroscopic sample, this approach is able to double purity results. Moreover, in order to take into account multiple spectroscopic observations in the same night, we propose a semisupervised batch-mode AL algorithm that selects a set of N = 5 most informative objects at each night. In comparison with the initial state using the traditional approach, our method achieves 2.3 times higher purity and comparable figure of merit results after only 180 d of observation, or 800 queries (73 per cent of the SNPCC spectroscopic sample size). Such results were obtained using the same amount of spectroscopic time necessary to observe the original SNPCC spectroscopic sample, showing that this type of strategy is feasible with current available spectroscopic resources. The code used in this work is available in the COINtoolbox. (AU)

Processo FAPESP: 09/54006-4 - Um cluster de computadores para o Departamento de Astronomia do IAG-USP e para o Núcleo de Astrofísica da UNICSUL
Beneficiário:Elisabete Maria de Gouveia Dal Pino
Linha de fomento: Auxílio à Pesquisa - Programa Equipamentos Multiusuários