dc.contributor.author | Salmeron, Antonio | |
dc.contributor.author | Langseth, Helge | |
dc.contributor.author | Masegosa, Andres | |
dc.contributor.author | Nielsen, Thomas D. | |
dc.date.accessioned | 2023-08-31T13:00:03Z | |
dc.date.available | 2023-08-31T13:00:03Z | |
dc.date.created | 2023-08-28T09:48:53Z | |
dc.date.issued | 2022 | |
dc.identifier.citation | Proceedings of Machine Learning Research (PMLR). 2022, 186 205-216. | en_US |
dc.identifier.issn | 2640-3498 | |
dc.identifier.uri | https://hdl.handle.net/11250/3086661 | |
dc.description.abstract | Mixtures of truncated basis functions (MoTBFs) are a popular tool within the context of hybrid Bayesian networks, mainly because they are compatible with e_cient probabilistic inference schemes. However, their standard parameterization allows the presence of negative mixture weights as well as non-normalized mixture terms, which prevents them from bene_ting from existing likelihood-based mixture estimation methods like the EM algorithm. Furthermore, the standard parameterization does not facilitate the de_nition of a Bayesian framework ideally allowing conjugate analysis. In this paper we show how MoTBFs can be reparameterized applying a strategy already used in the literature for Gaussian mixture models with negative terms. We exemplify how the new parameterization is compatible with the EM algorithm and conjugate analysis | en_US |
dc.language.iso | eng | en_US |
dc.publisher | MLResearchPress | en_US |
dc.title | A Reparameterization of Mixtures of Truncated Basis Functions and its Applications | en_US |
dc.title.alternative | A Reparameterization of Mixtures of Truncated Basis Functions and its Applications | en_US |
dc.type | Peer reviewed | en_US |
dc.type | Journal article | en_US |
dc.description.version | publishedVersion | en_US |
dc.rights.holder | Copyright © The authors and PMLR 2023 | en_US |
dc.source.pagenumber | 205-216 | en_US |
dc.source.volume | 186 | en_US |
dc.source.journal | Proceedings of Machine Learning Research (PMLR) | en_US |
dc.identifier.cristin | 2170068 | |
cristin.ispublished | true | |
cristin.fulltext | original | |
cristin.qualitycode | 1 | |