Combinando Métodos de Aprendizado Supervisionado para a Melhoria da Previsão do Redshift de Galáxias
DOI:
https://doi.org/10.5540/tema.2020.021.01.117Keywords:
Aprendizado de máquina, stacking, densidades condicionais, cosmologiaAbstract
Um problema fundamental em cosmologia é estimar redshifts de galáxias com base em dados fotométricos. Por exemplo a Sloan Digital Sky Survey (SDSS) já coletou dados fotométricos relativos a cerca de um bilhão de objetos para os quais é necessário estimar os respectivos redshifts. Tradicionalmente, essa tarefa é resolvida utilizando-se métodos de aprendizado de máquina. Neste trabalho, mostramos como métodos existentes podem ser combinados de forma a se obter estimativas ainda mais precisas para os redshifts de galáxias. Abordamos este problema sob duas óticas: (i) estimação da regressão do redshift y nas covariáveis fotométricas x, E[Y|x], e (ii) estimação da função densidade condicional f(y|x). Aplicamos as técnicas propostas para um banco de dados provenientes do SDSS e concluímos que as predições combinadas são de fato mais precisas que os métodos individuais.
References
Almosallam, Ibrahim A. ; Jarvis, Matt J. ; Roberts, Stephen J.: GPZ: nonstationary sparse Gaussian processes for heteroscedastic uncertainty estimation in photometric redshifts. In: Monthly Notices of the Royal Astronomical Society 462 (2016), Nr. 1, S. 726–739
Brammer, Gabriel B. ; Dokkum, Pieter G. ; Coppi, Paolo: EAZY: a fast,
public photometric redshift code. In: The Astrophysical Journal 686 (2008),
Nr. 2, S. 1503
Breiman, Leo: Random forests. In: Machine learning 45 (2001), Nr. 1, S.
–32
Chen, Tianqi ; Guestrin, Carlos: Xgboost: A scalable tree boosting system. In: Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining ACM, 2016, S. 785–794
Csabai, Istvan ; Budavari, Tamas ; Connolly, Andrew J. ; Szalay, Alexander S. ; Győry, Zsuzsanna ; Benitez, Narciso ; Annis, Jim ; Brinkmann, Jon ; Eisenstein, Daniel ; Fukugita, Masataka u. a.: The application of photometric redshifts to the SDSS early data release. In: The Astronomical Journal 125 (2003), Nr. 2, S. 580
Freeman, Peter E. ; Izbicki, Rafael ; Lee, Ann B.: A unified framework for constructing, tuning and assessing photometric redshift density estimates in a selection bias setting. In: Monthly Notices of the Royal Astronomical Society 468 (2017), Nr. 4, S. 4556–4565
Friedman, Jerome ; Hastie, Trevor ; Tibshirani, Robert: The elements of
statistical learning. Bd. 1. Springer series in statistics New York, 2001
Izbicki, R. ; Santos, T. M. d.: Machine Learning sob a ótica estatística. 2019 http://www.rizbicki.ufscar.br/sml
Izbicki, Rafael ; Lee, Ann B. u. a.: Converting high-dimensional regression to high-dimensional conditional density estimation. In: Electronic Journal of Statistics 11 (2017), Nr. 2, S. 2800–2831
Izbicki, Rafael ; Lee, Ann B. ; Freeman, Peter E.: Photo-z estimation: An example of nonparametric conditional density estimation under selection bias. In: The Annals of Applied Statistics 11 (2017), Nr. 2, S. 698–724
Sheldon, Erin S. ; Cunha, Carlos E. ; Mandelbaum, Rachel ; Brinkmann, J ; Weaver, Benjamin A.: Photometric redshift probability distributions for galaxies in the SDSS DR8. In: The Astrophysical Journal Supplement Series 201 (2012), Nr. 2, S. 32
Tibshirani, Robert: Regression shrinkage and selection via the lasso. In: Journal of the Royal Statistical Society: Series B (Methodological) 58 (1996), Nr. 1, S. 267–288
Turlach, Berwin A. ; Weingessel, Andreas: quadprog: Functions to solve quadratic programming problems. In: CRAN-Package quadprog (2007)
Wadadekar, Yogesh: Estimating photometric redshifts using support vector machines. In: Publications of the Astronomical Society of the Pacific 117 (2004), Nr. 827, S. 79
Wittman, D: What lies beneath: Using p(z) to reduce systematic photometric redshift errors. In: The Astrophysical Journal Letters 700 (2009), Nr. 2, S. L174
Yèche, Ch ; Petitjean, P ; Rich, J ; Aubourg, E ; Hamilton, J-Ch ;
Le Goff, J-M ; Paris, I ; Peirani, S ; Pichon, Ch ; Rollinde, E u. a.:
Artificial neural networks for quasar selection and photometric redshift determination. In: Astronomy & Astrophysics 523 (2010), S. A14
York, Donald G. ; Adelman, J ; Anderson Jr, John E. ; Anderson,
Scott F. ; Annis, James ; Bahcall, Neta A. ; Bakken, JA ; Barkhouser,
Robert ; Bastian, Steven ; Berman, Eileen u. a.: The sloan digital sky survey: Technical summary. In: The Astronomical Journal 120 (2000), Nr. 3, S. 1579
Zhang, Min-Ling ; Zhou, Zhi-Hua: ML-KNN: A lazy learning approach to multi-label learning. In: Pattern recognition 40 (2007), Nr. 7, S. 2038–2048
Zhou, Zhi-Hua: Ensemble methods: foundations and algorithms. CRC press, 2012
Downloads
Published
How to Cite
Issue
Section
License
Authors who publish in this journal agree to the following terms:
Authors retain copyright and grant the journal the right of first publication, with the work simultaneously licensed under the Creative Commons Attribution License that allows the sharing of the work with acknowledgment of authorship and initial publication in this journal.
Authors are authorized to assume additional contracts separately, for non-exclusive distribution of the version of the work published in this journal (eg, publish in an institutional repository or as a book chapter), with acknowledgment of authorship and initial publication in this journal.
Authors are allowed and encouraged to publish and distribute their work online (eg, in institutional repositories or on their personal page) at any point before or during the editorial process, as this can generate productive changes as well as increase impact and the citation of the published work (See The effect of open access).
This is an open access journal which means that all content is freely available without charge to the user or his/her institution. Users are allowed to read, download, copy, distribute, print, search, or link to the full texts of the articles, or use them for any other lawful purpose, without asking prior permission from the publisher or the
author. This is in accordance with the BOAI definition of open access
Intellectual Property
All the contents of this journal, except where otherwise noted, is licensed under a Creative Commons Attribution License under attribution BY.