论文信息 - NMF and PLSI: equivalence and a hybrid algorithm

NMF and PLSI: equivalence and a hybrid algorithm

In this paper, we show that PLSI and NMF optimize the same objective function, although PLSI and NMF are different algorithms as verified by experiments. In addition, we also propose a new hybrid method that runs PLSI and NMF alternatively to achieve better solutions.

Chris H. Q. Ding | Wei Peng | Tao Li

[1] Chris H. Q. Ding,et al. On the Equivalence of Nonnegative Matrix Factorization and Spectral Clustering , 2005, SDM.

[2] Thomas Hofmann,et al. Probabilistic Latent Semantic Analysis , 1999, UAI.

[3] Éric Gaussier,et al. Relation between PLSA and NMF and implications , 2005, SIGIR '05.

[4] Tao Li,et al. Document clustering via adaptive subspace iteration , 2004, SIGIR '04.

[5] H. Sebastian Seung,et al. Algorithms for Non-negative Matrix Factorization , 2000, NIPS.

[6] Chris H. Q. Ding,et al. Nonnegative Matrix Factorization and Probabilistic Latent Semantic Indexing: Equivalence Chi-Square Statistic, and a Hybrid Method , 2006, AAAI.