Topic discovery and tracking

from Wikipedia, the free encyclopedia

The research field of the topics of discovery and tracking ( Topic Detection and Tracking , TDT, and event-based information organization ) deals with the development of technologies which enable messages from television , the Internet or radio capture and these then into individual Separate messages and classify them into specific subject areas or a subject area. They are used, for example, by the news service Google News .

The research was driven by DARPA with the aim of making it easier for news analysts to deal with the growing flood of information.

In contrast to the traditional task of information retrieval , there is no clear need for information on the part of the searcher, but the aim is to identify new topics.

The problem is broken down into five tasks

  1. Segmentation: separation of texts into individual messages
  2. Topic discovery: Identification of new topics and grouping of reports by topic
  3. Cluster detection: Classification of incoming reports according to topic
  4. Topic tracking: Finding further reports on a topic
  5. Link detection: Determination of whether two randomly selected messages deal with a common topic

Information retrieval, text mining and computational linguistics techniques are used to accomplish the tasks .

Individual evidence

  1. Topic discovery and tracking ( memento of the original from July 25, 2014 in the Internet Archive ) Info: The archive link was inserted automatically and has not yet been checked. Please check the original and archive link according to the instructions and then remove this notice. , Publication by Wolfgang G. Stock  @1@ 2Template: Webachiv / IABot / www.phil-fak.uni-duesseldorf.de
  2. ^ Topic Detection and Tracking . Event-based information organization. In: James Allan (Ed.): The Information Retrieval Series . Vol. 12. Springer, 2002, ISBN 978-0-7923-7664-4 , pp. 3 f . ( limited preview in Google Book search).
  3. Juha, Makkonen, Helena Ahonen-Myka, Marko Salmenkivi: Simple Semantics in Topic Detection and Tracking . In: Information Retrieval . Vol. 7, No. 3 . Springer, 2004, p. 347-368 , doi : 10.1023 / B: INRT.0000011210.12953.86 .

literature

  • James Allan (Ed.): Topic Detection and Tracking. Event-based information organization. Kluwer, Boston MA et al. 2002, ISBN 0-7923-7664-1 ( Kluwer International Series on Information Retrieval 12).