eprintid: 10180942 rev_number: 6 eprint_status: archive userid: 699 dir: disk0/10/18/09/42 datestamp: 2023-11-10 13:58:33 lastmod: 2023-11-10 13:58:33 status_changed: 2023-11-10 13:58:33 type: proceedings_section metadata_visibility: show sword_depositor: 699 creators_name: Sellier, J creators_name: Dellaportas, P title: Bayesian online change point detection with Hilbert space approximate Student-t process ispublished: pub divisions: UCL divisions: B04 divisions: C06 divisions: F61 note: © The Authors 2023. Original content in this paper is licensed under the terms of the Creative Commons Attribution 4.0 International (CC BY 4.0) Licence (https://creativecommons.org/licenses/by/4.0/). abstract: In this paper, we introduce a variant of Bayesian online change point detection with a reduced-rank Student-t process (TP) and dependent Student-t noise, as a nonparametric time series model. Our method builds and improves upon the state-of-the-art Gaussian process (GP) change point model benchmark of Saatçi et al. (2010). The Student-t process generalizes the concept of a GP and hence yields a more flexible alternative. Additionally, unlike a GP, the predictive variance explicitly depends on the training observations, while the use of an entangled Student-t noise model preserves analytical tractability. Our approach also uses a Hilbert space reduced-rank representation of the TP kernel, derived from an eigenfunction expansion of the Laplace operator (Solin & Särkkä, 2020), to alleviate its computational complexity. Improvements in prediction and training time are demonstrated with real-world data sets. date: 2023 date_type: published publisher: PMLR (Proceedings of Machine Learning Research) official_url: https://proceedings.mlr.press/v202/sellier23a.html oa_status: green full_text_type: pub language: eng primo: open primo_central: open_green verified: verified_manual elements_id: 2106060 lyricists_name: Dellaportas, Petros lyricists_id: PDELL40 actors_name: Flynn, Bernadette actors_id: BFFLY94 actors_role: owner full_text_status: public pres_type: paper publication: Proceedings of Machine Learning Research volume: 202 pagerange: 30553-30569 issn: 2640-3498 book_title: Proceedings of the 40th International Conference on Machine Learning citation: Sellier, J; Dellaportas, P; (2023) Bayesian online change point detection with Hilbert space approximate Student-t process. In: Proceedings of the 40th International Conference on Machine Learning. (pp. pp. 30553-30569). PMLR (Proceedings of Machine Learning Research) Green open access document_url: https://discovery.ucl.ac.uk/id/eprint/10180942/1/sellier23a.pdf