Relation between PLSA and NMF and implications

Abstract

Non-negative Matrix Factorization (NMF, [5]) and Probabilistic Latent Semantic Analysis (PLSA, [4]) have been successfully applied to a number of text analysis tasks such as document clustering. Despite their different inspirations, both methods are instances of multinomial PCA [1]. We further explore this relationship and first show that PLSA solves the problem of NMF with KL divergence, and then explore the implications of this relationship.

References

Page 1

	Year	Citations
Learning the parts of objects by non-negative matrix factorization Daniel D. Lee, H. Sebastian Seung Nature Non-negative Matrix FactorizationMachine VisionImage AnalysisData ScienceMachine Learning	1999	13.8K
ALGORITHMS FOR NON-NEGATIVE MATRIX FACTORIZATION D Seung, Lin-Wen Lee EngineeringMatrix FactorizationData MiningKnowledge DiscoveryMultilinear Subspace Learning	2001	4.8K
Probabilistic Latent Semantic Analysis Thomas Hofmann arXiv (Cornell University) EngineeringSemanticsCorpus LinguisticsText MiningWord Embeddings	2013	2.1K
Document clustering based on non-negative matrix factorization Wei Xu, Xin Liu, Yihong Gong Natural Language ProcessingDocument ClusteringNon-negative Matrix FactorizationEngineeringInformation Retrieval	2003	1.8K
Aligning words using matrix factorisation Cyril Goutte, Kenji Yamada, Éric Gaussier Natural Language ProcessingComputer-assisted TranslationText NormalizationEngineeringMatrix Factorization	2004	27

Page 1