Nonparametric bayesian nonnegative matrix factorization

Hong Bo Xie*, Caoyuan Li, Kerrie Mengersen, Shuliang Wang, Richard Yi Da Xu

*此作品的通讯作者

科研成果: 书/报告/会议事项章节会议稿件同行评审

摘要

Nonnegative Matrix Factorization (NMF) is an important tool in machine learning for blind source separation and latent factor extraction. Most of existing NMF algorithms assume a specific noise kernel, which is insufficient to deal with complex noise in real scenarios. In this study, we present a hierarchical nonparametric nonnegative matrix factorization (NPNMF) model in which the Gaussian mixture model is used to approximate the complex noise distribution. The model is cast in the nonparametric Bayesian framework by using Dirichlet process mixture to infer the necessary number of Gaussian components. We derive a mean-field variational inference algorithm for the proposed nonparametric Bayesian model. Experimental results on both synthetic data and electroencephalogram (EEG) demonstrate that NPNMF performs better in extracting the latent nonnegative factors in comparison with state-of-the-art methods.

源语言英语
主期刊名Modeling Decisions for Artificial Intelligence - 17th International Conference, MDAI 2020, Proceedings
编辑Vicenc Torra, Yasuo Narukawa, Jordi Nin, Núria Agell
出版商Springer
132-141
页数10
ISBN(印刷版)9783030575236
DOI
出版状态已出版 - 2020
活动17th International Conference on Modeling Decisions for Artificial Intelligence, MDAI 2020 - Sant Cugat del Vallès, 西班牙
期限: 2 9月 20204 9月 2020

出版系列

姓名Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
12256 LNAI
ISSN(印刷版)0302-9743
ISSN(电子版)1611-3349

会议

会议17th International Conference on Modeling Decisions for Artificial Intelligence, MDAI 2020
国家/地区西班牙
Sant Cugat del Vallès
时期2/09/204/09/20

指纹

探究 'Nonparametric bayesian nonnegative matrix factorization' 的科研主题。它们共同构成独一无二的指纹。

引用此