Identifying and propagating contextually appropriate deep-topics amongst collaborating web-users

Jeremy Hall, Yasushi Kiyoki

Research output: Chapter in Book/Report/Conference proceedingChapter

1 Citation (Scopus)


This paper describes a method for discovering URLs with contextually relevant deep-topics, and then propagating such information to collaborating users lacking such information. When a user is knowledgeable about a subject, their reasons for frequently browsing a URL extend beyond the fact that it is merely related to said subject. This paper's method includes an algorithm for discovering the surface-topic of a URL, and the underlying deep-topic that a user is truly interested in with respect to a given URL. The deep-topic extraction process works by using URLs linked together through a user's behavioral browsing patterns in order to discover the surface or group-topic of surrounding URLs, and then subtracting those topics to discover hidden deeper topics. This paper describes the three parts of the method: Information Extraction, Propagation, and Verification & Integration, which together form a method with high levels of parallelism due to its distributed and independent nature. This paper also discusses concrete usage-scenarios for the included method, and data structures which would support the implementation of this paper's method.

Original languageEnglish
Title of host publicationInformation Modelling and Knowledge Bases XXV
EditorsTakehiro Tokuda, Yasushi Kiyoki, Hannu Jaakkola, Naofumi Yoshida
Number of pages12
Publication statusPublished - 2014 Mar 3

Publication series

NameFrontiers in Artificial Intelligence and Applications
ISSN (Print)0922-6389


  • collaboration
  • data mining
  • knowledge discovery
  • knowledge sharing
  • web behavior

ASJC Scopus subject areas

  • Artificial Intelligence


Dive into the research topics of 'Identifying and propagating contextually appropriate deep-topics amongst collaborating web-users'. Together they form a unique fingerprint.

Cite this