Citationsanalyse & Automatisk Klassifikation

Studenteropgave: Bachelorprojekt

  • Asbjørn Dahl
  • Heidi Larsen
6. sem, Bachelor i Biblioteks- og informationsvidenskab
Introduction: This dissertation is a feasibility study of automatic classification of citations – in theory and practice. It is a critical assessment of the potential use of automatic classification in the different application areas of bibliometrics.
Method: A experiment was performed in automatic classification of 1213 references in 41 articles from the journal Information Research. This serves as a proof of concept, in order to evaluate the practical foundation for performing automated classifications of citations using machine learning.
Analysis: We discuss the theoretical use of results from automatic classification of citations in the different application areas of bibliometrics. We discuss the validity of such results from the perspective of classification theory.
Results: There are non-trivial problems with assigning an unbiased classification scheme as the foundation for automatic classification. This is a problem in the area of research evaluation. There is potential in using automatic classification to improve information retrieval in citation indexes and to study science as a social network in greater detail.
Conclusion: We conclude that there are several areas of citation analysis in which automatic classification of citations could prove beneficial. But there are still a need for further technical developments in the processing of text and preliminary studies in classification schemes for citation types.
Udgivelsesdato27 maj 2013
Antal sider38
Udgivende institutionDet Informationsvidenskabelige Akademi


  • bibliometri, citationsanalyse, maskinlæring, klassifikationsteori