eprintid: 10064561 rev_number: 16 eprint_status: archive userid: 608 dir: disk0/10/06/45/61 datestamp: 2018-12-20 11:49:14 lastmod: 2021-09-23 22:31:47 status_changed: 2018-12-20 11:49:14 type: article metadata_visibility: show creators_name: Li, L creators_name: Guedj, B creators_name: Loustau, S title: A quasi-Bayesian perspective to online clustering ispublished: pub divisions: UCL divisions: B04 divisions: C05 divisions: F48 keywords: Online clustering, quasi-Bayesian learning, minimax regret bounds, reversible jump Markov chain Monte Carlo note: This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. abstract: When faced with high frequency streams of data, clustering raises theoretical and algorithmic pitfalls. We introduce a new and adaptive online clustering algorithm relying on a quasi-Bayesian approach, with a dynamic (i.e., time-dependent) estimation of the (unknown and changing) number of clusters. We prove that our approach is supported by minimax regret bounds. We also provide an RJMCMC-flavored implementation (called PACBO, see https://cran.r-project.org/web/packages/PACBO/index.html) for which we give a convergence guarantee. Finally, numerical experiments illustrate the potential of our procedure. date: 2018-01 date_type: published official_url: https://doi.org/10.1214/18-EJS1479 oa_status: green full_text_type: pub language: eng primo: open primo_central: open_green verified: verified_manual elements_id: 1605922 doi: 10.1214/18-EJS1479 lyricists_name: Guedj, Benjamin lyricists_id: BGUED94 actors_name: Flynn, Bernadette actors_id: BFFLY94 actors_role: owner full_text_status: public publication: Electronic Journal of Statistics volume: 12 number: 2 pagerange: 3071-3113 issn: 1935-7524 citation: Li, L; Guedj, B; Loustau, S; (2018) A quasi-Bayesian perspective to online clustering. Electronic Journal of Statistics , 12 (2) pp. 3071-3113. 10.1214/18-EJS1479 <https://doi.org/10.1214/18-EJS1479>. Green open access document_url: https://discovery.ucl.ac.uk/id/eprint/10064561/1/euclid.ejs.1537430425.pdf