A machine learning approach to galaxy properties: Joint redshift - stellar mass probability distributions with Random Forest
A machine learning approach to galaxy properties: Joint redshift - stellar mass probability distributions with Random Forest
复制标题
DOI：
10.1093/mnras/stab164
复制
发表时间：
2020-12
期刊：
影响因子：
--
通讯作者：
S. Mucesh;W. Hartley;A. Palmese;O. Lahav;L. Whiteway;A. Amon;K. Bechtol;G. Bernstein;A. Rosell;M. Kind;A. Choi;K. Eckert;S. Everett;D. Gruen;R. Gruendl;I. Harrison;E. Huff;N. Kuropatkin;I. Sevilla-Noarbe;E. Sheldon;B. Yanny;M. Aguena;S. Allam;D. Bacon;E. Bertin;S. Bhargava;D. Brooks;J. Carretero;F. Castander;C. Conselice;M. Costanzi;M. Crocce;L. Costa;M. Pereira;J. DeVicente;S. Desai;H. Diehl;A. Drlica-Wagner;A. Evrard;I. Ferrero;B. Flaugher;P. Fosalba;J. Frieman;J. García-Bellido;E. Gaztañaga;D. Gerdes;J. Gschwend;G. Gutiérrez;S. Hinton;D. Hollowood;K. Honscheid;D. James;K. Kuehn;M. Lima;H. Lin;M. Maia;Peter Melchior;F. Menanteau;R. Miquel;R. Morgan;F. Paz-Chinchón;A. Plazas;E. Sánchez;V. Scarpine;M. Schubnell;S. Serrano;M. Smith;E. Suchyta;G. Tarlé;D. Thomas;C. To;T. Varga;R. Wilkinson
中科院分区：
文献类型：
--
作者：
S. Mucesh;W. Hartley;A. Palmese;O. Lahav;L. Whiteway;A. Amon;K. Bechtol;G. Bernstein;A. Rosell;M. Kind;A. Choi;K. Eckert;S. Everett;D. Gruen;R. Gruendl;I. Harrison;E. Huff;N. Kuropatkin;I. Sevilla-Noarbe;E. Sheldon;B. Yanny;M. Aguena;S. Allam;D. Bacon;E. Bertin;S. Bhargava;D. Brooks;J. Carretero;F. Castander;C. Conselice;M. Costanzi;M. Crocce;L. Costa;M. Pereira;J. DeVicente;S. Desai;H. Diehl;A. Drlica-Wagner;A. Evrard;I. Ferrero;B. Flaugher;P. Fosalba;J. Frieman;J. García-Bellido;E. Gaztañaga;D. Gerdes;J. Gschwend;G. Gutiérrez;S. Hinton;D. Hollowood;K. Honscheid;D. James;K. Kuehn;M. Lima;H. Lin;M. Maia;Peter Melchior;F. Menanteau;R. Miquel;R. Morgan;F. Paz-Chinchón;A. Plazas;E. Sánchez;V. Scarpine;M. Schubnell;S. Serrano;M. Smith;E. Suchyta;G. Tarlé;D. Thomas;C. To;T. Varga;R. Wilkinson
我们证明，即使可用的光度波段很少，也可以使用随机森林（RF）机器学习（ML）算法获得高精度的联合红移-恒星质量概率分布函数（PDF）。例如，我们使用暗能量巡天 (DES) 以及 COSMOS2015 红移和恒星质量目录。我们构建了两个 ML 模型：一个包含灰波段的深度光度测量，第二个反映主要 DES 调查中存在的光度散射，并在每种情况下精心构建了代表性训练数据。我们利用 copula 概率积分变换和 Kendall 分布函数以及它们的单变量对应函数来验证边缘，从而验证 10 699 个测试星系的联合 PDF。根据模板拟合代码风笛的基本设置进行基准测试，我们基于 ML 的方法在所有预定义的性能指标上都优于模板拟合。除了准确性之外，RF 的速度也非常快，能够使用消费类计算机硬件在不到 6 分钟的时间内计算出一百万个星系的联合 PDF。这样的速度使得 PDF 能够在分析代码中实时生成，从而解决潜在的存储问题。作为这项工作的一部分，我们开发了 galpro1，这是一个高度直观且高效的 python 包，用于快速生成多元 PDF。 galpro 已被记录并可供研究人员在其宇宙学和星系演化研究中使用。
We demonstrate that highly accurate joint redshift–stellar mass probability distribution functions (PDFs) can be obtained using the Random Forest (RF) machine learning (ML) algorithm, even with few photometric bands available. As an example, we use the Dark Energy Survey (DES), combined with the COSMOS2015 catalogue for redshifts and stellar masses. We build two ML models: one containing deep photometry in the griz bands, and the second reflecting the photometric scatter present in the main DES survey, with carefully constructed representative training data in each case. We validate our joint PDFs for 10 699 test galaxies by utilizing the copula probability integral transform and the Kendall distribution function, and their univariate counterparts to validate the marginals. Benchmarked against a basic set-up of the template-fitting code bagpipes, our ML-based method outperforms template fitting on all of our predefined performance metrics. In addition to accuracy, the RF is extremely fast, able to compute joint PDFs for a million galaxies in just under 6 min with consumer computer hardware. Such speed enables PDFs to be derived in real time within analysis codes, solving potential storage issues. As part of this work we have developed galpro1, a highly intuitive and efficient python package to rapidly generate multivariate PDFs on-the-fly. galpro is documented and available for researchers to use in their cosmology and galaxy evolution studies.