Discovering hierarchical topic evolution in time‐stamped documents
- 5 June 2015
- journal article
- research article
- Published by Wiley in Journal of the Association for Information Science and Technology
- Vol. 67 (4) , 915-927
Abstract
The objective of this paper is to propose a hierarchical topic evolution model (HTEM) that can organize time‐varying topics in a hierarchy and discover their evolutions with multiple timescales. In the proposed HTEM, topics near the root of the hierarchy are more abstract and also evolve in the longer timescales than those near the leaves. To achieve this goal, the distance‐dependent Chinese restaurant process (ddCRP) is extended to a new nested process that is able to simultaneously model the dependencies among data and the relationship between clusters. The HTEM is proposed based on the new process for time‐stamped documents, in which the timestamp is utilized to measure the dependencies among documents. Moreover, an efficient Gibbs sampler is developed for the proposed HTEM. Our experimental results on two popular real‐world data sets verify that the proposed HTEM can capture coherent topics and discover their hierarchical evolutions. It also outperforms the baseline model in terms of likelihood on held‐out data.Keywords
Funding Information
- National High Technology Research and Development Program of China (2012AA011005)
This publication has 22 references indexed in Scilit:
- A nonparametric mixture model for topic modeling over time2013
- Trending Twitter topics in English: An international comparisonJournal of the American Society for Information Science and Technology, 2012
- The dynamic features of Delicious, Flickr, and YouTubeJournal of the American Society for Information Science and Technology, 2011
- The shifting sands of disciplinary development: Analyzing North American Library and Information Science dissertations using latent Dirichlet allocationJournal of the American Society for Information Science and Technology, 2010
- New event detection and topic tracking in TurkishJournal of the American Society for Information Science and Technology, 2010
- The nested chinese restaurant process and bayesian nonparametric inference of topic hierarchiesJournal of the ACM, 2010
- Dynamic Non-Parametric Mixture Models and The Recurrent Chinese Restaurant Process: with Applications to Evolutionary Clustering2008
- A correlated topic model of ScienceThe Annals of Applied Statistics, 2007
- Finding scientific topicsProceedings of the National Academy of Sciences of the United States of America, 2004
- 10.1162/jmlr.2003.3.4-5.993Applied Physics Letters, 2000