Computer Science > Machine Learning
[Submitted on 18 May 2012 (this version), latest version 25 May 2012 (v2)]
Title:Theory of Dependent Hierarchical Normalized Random Measures
View PDFAbstract:This paper presents theory for Normalized Random Measures (NRMs), Normalized Generalized Gammas (NGGs), a particular kind of NRM, and Dependent Hierarchical Normalized Random Measures which have been used for time-dependent topic modelling. In this paper, we first introduce in some mathematical background of completely random measures (CRM) and their constructions from Poisson processes, as well as dependency operators in Poisson processes and the corresponding CRMs. The Normalized Generalised Gamma (NGG) is introduced. Slice sampling is also introduced to do the posterior sampling of normalized random measures. Operators on CRMs and NRMs are then given. Posterior inference on the NGGs is presented and it is shown that marginalizing the mass parameter of the NGG yields a Poisson-Dirichlet distribution. Finally, we give dependency results when applying these operators to NRMs. Proofs of related Lemmas and Theorems are given in the Appendix.
Submission history
From: Wray Buntine [view email][v1] Fri, 18 May 2012 13:56:17 UTC (41 KB)
[v2] Fri, 25 May 2012 05:32:57 UTC (44 KB)
Current browse context:
cs.LG
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
IArxiv Recommender
(What is IArxiv?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.