Clustering and mapping related news about violence events on their time-lines

Syed Toufeeq Ahmed, Sukru Tikves, Hasan Davulcu

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

Abstract

Keeping track of news stories and events as they progress can be a tedious job, but as every day routine most of the web users read and follow many stories and events in news. If an analyst in her area has to follow and map all these according to the time-line they happen, the task quickly becomes overwhelming. We present an online tool which attempts to ease the analyst's task of finding all news articles about an event, and sorting and mapping them on a time-line. We implemented an incremental clustering algorithm working on real-time incoming news, experimenting with different feature sets, including named entities and sentence overlap methods. We evaluated these approaches using Document Understand Conference (DUC) datasets.

Original languageEnglish (US)
Title of host publicationISI 2010 - 2010 IEEE International Conference on Intelligence and Security Informatics
Subtitle of host publicationPublic Safety and Security
Number of pages1
DOIs
StatePublished - Jul 26 2010
Event2010 IEEE International Conference on Intelligence and Security Informatics: Public Safety and Security, ISI 2010 - Vancouver, BC, Canada
Duration: May 23 2010May 26 2010

Publication series

NameISI 2010 - 2010 IEEE International Conference on Intelligence and Security Informatics: Public Safety and Security

Other

Other2010 IEEE International Conference on Intelligence and Security Informatics: Public Safety and Security, ISI 2010
CountryCanada
CityVancouver, BC
Period5/23/105/26/10

ASJC Scopus subject areas

  • Artificial Intelligence
  • Information Systems
  • Safety, Risk, Reliability and Quality

Fingerprint Dive into the research topics of 'Clustering and mapping related news about violence events on their time-lines'. Together they form a unique fingerprint.

Cite this