Event detection with spatial latent Dirichlet allocation

Chi Chun Pan, Prasenjit Mitra

Research output: Chapter in Book/Report/Conference proceedingConference contribution

34 Citations (Scopus)

Abstract

A large number of news articles are generated every day on the Web. Automatically identifying events from a large document collection is a challenging problem. In this paper, we propose two event detection approaches using generative models. We combine the popular LDA model with temporal segmentation and spatial clustering. In addition, we adapt an image segmentation model, SLDA, for spatial-temporal event detection on text. The results of our experiments show that both approaches outperform the traditional content-based clustering approaches on our datasets.

Original languageEnglish
Title of host publicationProceedings of the ACM/IEEE Joint Conference on Digital Libraries
Pages349-358
Number of pages10
DOIs
Publication statusPublished - 2011
Externally publishedYes
Event11th Annual International ACM/IEEE Joint Conference on Digital Libraries, JCDL'11 - Ottawa, ON
Duration: 13 Jun 201117 Jun 2011

Other

Other11th Annual International ACM/IEEE Joint Conference on Digital Libraries, JCDL'11
CityOttawa, ON
Period13/6/1117/6/11

    Fingerprint

Keywords

  • event detection
  • spatial segmentation
  • temporal clustering

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Pan, C. C., & Mitra, P. (2011). Event detection with spatial latent Dirichlet allocation. In Proceedings of the ACM/IEEE Joint Conference on Digital Libraries (pp. 349-358) https://doi.org/10.1145/1998076.1998141