Selecting the Links in BisoNets Generated from Document Collections

  • Marc Segond
  • Christian Borgelt
Conference paper

DOI: 10.1007/978-3-642-13062-5_19

Part of the Lecture Notes in Computer Science book series (LNCS, volume 6065)
Cite this paper as:
Segond M., Borgelt C. (2010) Selecting the Links in BisoNets Generated from Document Collections. In: Cohen P.R., Adams N.M., Berthold M.R. (eds) Advances in Intelligent Data Analysis IX. IDA 2010. Lecture Notes in Computer Science, vol 6065. Springer, Berlin, Heidelberg

Abstract

According to Koestler, the notion of a bisociation denotes a connection between pieces of information from habitually separated domains or categories. In this paper, we consider a methodology to find such bisociations using a network representation of knowledge, which is called a BisoNet, because it promises to contain bisociations. In a first step, we consider how to create BisoNets from several textual databases taken from different domains using simple text-mining techniques. To achieve this, we introduce a procedure to link nodes of a BisoNet and to endow such links with weights, which is based on a new measure for comparing text frequency vectors. In a second step, we try to rediscover known bisociations, which were originally found by a human domain expert, namely indirect relations between migraine and magnesium as they are hidden in medical research articles published before 1987. We observe that these bisociations are easily rediscovered by simply following the strongest links. Future work includes extending our methods to non-textual data, improving the similarity measure, and applying more sophisticated graph mining methods.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Copyright information

© Springer-Verlag Berlin Heidelberg 2010

Authors and Affiliations

  • Marc Segond
    • 1
  • Christian Borgelt
    • 1
  1. 1.European Center for Soft ComputingMieres (Asturias)Spain

Personalised recommendations