跳转到内容

PyMC

维基百科,自由的百科全书
(重定向自PyMC3
PyMC
原作者PyMC开发团队
首次发布2013年5月4日 (2013-05-04)
当前版本5.16.2(2024年7月11日 (2024-07-11)
源代码库https://github.com/pymc-devs/pymc
编程语言Python
操作系统类Unix, Mac OS X, Microsoft Windows
平台Intel x86 – 32-bit, x64
类型统计包英语List of statistical software
许可协议 Apache License, Version 2.0
网站www.pymc.io

PyMC(曾叫做PyMC3[1])是一个Python包,用于贝叶斯统计建模概率机器学习,它聚焦于高级马尔可夫链蒙特卡洛法和变分拟合算法[2][3][4]

概述

[编辑]

PyMC曾经叫做PyMC3,不同于早先的使用Fortran扩展进行计算的PyMC2,它依靠Theano来进行自动微分、计算优化和动态C语言编译[3][5]。从版本3.8开始PyMC依据ArviZ英语ArviZ来进行数据可视化贝叶斯推断探索分析英语Exploratory data analysis[6]。PyMC和Stan英语Stan (software)是两个最流行的概率编程工具[7]

PyMC是开源项目,由社区开发并在财务上得到NumFocus赞助[8]。PyMC已经在很多领域中被用于解决推断问题,包括天文学[9][10]流行病学[11][12]分子生物学[13]晶体学[14][15]化学[16]生态学[17][18]心理学[19]

Theano于2017年宣布计划停止开发之后[20],PyMC团队曾评估采用TensorFlow Probability[21]作为计算后端[22],但是在2020年接管Theano的开发[23]。在2021年1月绝大部份的Theano代码基被重新建造,并增加了通过JAXNumba的编译,修订后的这个计算后端以新名字Aesara发行。PyMC团队在2021年6月将PyMC3更名为PyMC[1]。2022年11月28日,PyMC团队宣布采用从Aesara计划分叉出PyTensor[24]

推论引擎

[编辑]

PyMC实现了不基于梯度的和基于梯度的马尔可夫链蒙特卡洛(MCMC)算法用于贝叶斯推断和随机(Stochastic英语Stochastic),基于梯度的变分贝叶斯方法用于近似贝叶斯推断。

参见

[编辑]

引用

[编辑]
  1. ^ 1.0 1.1 PyMC Timeline. PyMC Timeline. [2021-01-20]. (原始内容存档于2018-05-20). 
  2. ^ Salvatier J, Wiecki TV, Fonnesbeck C. (2016) Probabilistic programming in Python using PyMC3. PeerJ Computer Science 2:e55 https://doi.org/10.7717/peerj-cs.55
  3. ^ 3.0 3.1 Martin, Osvaldo. Bayesian Analysis with Python. Packt Publishing Ltd. 2016: 31–60 [16 September 2017]. ISBN 9781785889851 (英语). 
  4. ^ Davidson-Pilon, Cameron. Bayesian Methods for Hackers: Probabilistic Programming and Bayesian Inference. Addison-Wesley Professional. 2015-09-30. ISBN 9780133902921 (英语). 
  5. ^ Hilpisch, Yves. Python for Finance: Analyze Big Financial Data. O'Reilly Media, Inc. 2014-12-11. ISBN 9781491945391 (英语). 
  6. ^ ArviZ — Exploratory analysis of Bayesian models. [2023-09-21]. (原始内容存档于2023-10-11). 
  7. ^ The Algorithms Behind Probabilistic Programming. [2017-03-10]. (原始内容存档于2021-01-29). 
  8. ^ NumFOCUS Announces New Fiscally Sponsored Project: PyMC3. NumFOCUS | Open Code = Better Science. [2017-03-10]. (原始内容存档于2017-09-21). 
  9. ^ Greiner, J.; Burgess, J. M.; Savchenko, V.; Yu, H.-F. On the Fermi-GBM Event 0.4 s after GW150914. The Astrophysical Journal Letters. 2016, 827 (2): L38. Bibcode:2016ApJ...827L..38G. ISSN 2041-8205. arXiv:1606.00314可免费查阅. doi:10.3847/2041-8205/827/2/L38 (英语). 
  10. ^ Hilbe, Joseph M.; Souza, Rafael S. de; Ishida, Emille E. O. Bayesian Models for Astrophysical Data: Using R, JAGS, Python, and Stan. Cambridge University Press. 2017-04-30 [2021-01-20]. ISBN 9781108210744. (原始内容存档于2021-02-03) (英语). 
  11. ^ Brauner, Jan M.; Mindermann, Sören; Sharma, Mrinank; Johnston, David; Salvatier, John; Gavenčiak, Tom; Stephenson, Anna B.; Leech, Gavin; Altman, George; Mikulik, Vladimir; Norman, Alexander John; Monrad, Joshua Teperowski; Besiroglu, Tamay; Ge, Hong; Hartwick, Meghan A.; Teh, Yee Whye; Chindelevitch, Leonid; Gal, Yarin; Kulveit, Jan. Inferring the effectiveness of government interventions against COVID-19. Science. 2020-12-15 [2021-01-20]. doi:10.1126/science.abd9338. (原始内容存档于2021-02-07). 
  12. ^ Systrom, Kevin; Vladek, Thomas; Krieger, Mike. Rt.live Github repository. Rt.live. [10 January 2021]. (原始内容存档于2021-01-06). 
  13. ^ Wagner, Stacey D.; Struck, Adam J.; Gupta, Riti; Farnsworth, Dylan R.; Mahady, Amy E.; Eichinger, Katy; Thornton, Charles A.; Wang, Eric T.; Berglund, J. Andrew. Dose-Dependent Regulation of Alternative Splicing by MBNL Proteins Reveals Biomarkers for Myotonic Dystrophy. PLOS Genetics. 2016-09-28, 12 (9): e1006316. ISSN 1553-7404. PMC 5082313可免费查阅. PMID 27681373. doi:10.1371/journal.pgen.1006316. 
  14. ^ Sharma, Amit; Johansson, Linda; Dunevall, Elin; Wahlgren, Weixiao Y.; Neutze, Richard; Katona, Gergely. Asymmetry in serial femtosecond crystallography data. Acta Crystallographica Section A. 2017-03-01, 73 (2): 93–101. ISSN 2053-2733. PMC 5332129可免费查阅. PMID 28248658. doi:10.1107/s2053273316018696 (英语). 
  15. ^ Katona, Gergely; Garcia-Bonete, Maria-Jose; Lundholm, Ida. Estimating the difference between structure-factor amplitudes using multivariate Bayesian inference. Acta Crystallographica Section A. 2016-05-01, 72 (3): 406–411. ISSN 2053-2733. PMC 4850660可免费查阅. PMID 27126118. doi:10.1107/S2053273316003430 (英语). 
  16. ^ Garay, Pablo G.; Martin, Osvaldo A.; Scheraga, Harold A.; Vila, Jorge A. Detection of methylation, acetylation and glycosylation of protein residues by monitoring13C chemical-shift changes: A quantum-chemical study. PeerJ. 2016-07-21, 4: e2253. ISSN 2167-8359. PMC 4963218可免费查阅. PMID 27547559. doi:10.7717/peerj.2253 (英语). 
  17. ^ Wang, Yan; Huang, Hong; Huang, Lida; Ristic, Branko. Evaluation of Bayesian source estimation methods with Prairie Grass observations and Gaussian plume model: A comparison of likelihood functions and distance measures. Atmospheric Environment. 2017, 152: 519–530. Bibcode:2017AtmEn.152..519W. doi:10.1016/j.atmosenv.2017.01.014. 
  18. ^ MacNeil, M. Aaron; Chong-Seng, Karen M.; Pratchett, Deborah J.; Thompson, Casssandra A.; Messmer, Vanessa; Pratchett, Morgan S. Age and Growth of An Outbreaking Acanthaster cf. solaris Population within the Great Barrier Reef. Diversity. 2017-03-14, 9 (1): 18. doi:10.3390/d9010018 (英语). 
  19. ^ Tünnermann, Jan; Scharlau, Ingrid. Peripheral Visual Cues: Their Fate in Processing and Effects on Attention and Temporal-Order Perception. Frontiers in Psychology. 2016, 7. ISSN 1664-1078. PMC 5052275可免费查阅. PMID 27766086. doi:10.3389/fpsyg.2016.01442 (英语). 
  20. ^ Lamblin, Pascal. MILA and the future of Theano. theano-users (邮件列表). 28 September 2017 [28 September 2017]. (原始内容存档于2011-01-22). 
  21. ^ TensorFlow Probability is a library for probabilistic reasoning and statistical analysis. [2022-08-31]. (原始内容存档于2022-09-04). 
  22. ^ Developers, PyMC. Theano, TensorFlow and the Future of PyMC. PyMC Developers. 2018-05-17 [2019-01-25]. (原始内容存档于2020-08-06). 
  23. ^ The Future of PyMC3, or: Theano is Dead, Long Live Theano. PyMC Developers. [10 January 2021]. (原始内容存档于2021-01-15). 
  24. ^ PyMC forked Aesara to PyTensor. [2023-08-17]. (原始内容存档于2023-07-18). 
  25. ^ Hoffman, Matthew D.; Gelman, Andrew. The No-U-Turn Sampler: Adaptively Setting Path Lengths in Hamiltonian Monte Carlo. Journal of Machine Learning Research. April 2014, 15: pp. 1593–1623 [2021-01-20]. (原始内容存档于2020-08-11). 
  26. ^ Kucukelbir, Alp; Ranganath, Rajesh; Blei, David M. Automatic Variational Inference in Stan 1506 (3431). June 2015. Bibcode:2015arXiv150603431K. arXiv:1506.03431可免费查阅. 

延伸阅读

[编辑]

外部链接

[编辑]