• Cluster-based polyrepresentation as science modelling approach for information retrieval

      Abbasi, Muhammad Kamran; Frommholz, Ingo (Springer Verlag, 2015)
      The increasing number of publications make searching and accessing the produced literature a challenging task. A recent development in bibliographic databases is to use advanced information retrieval techniques in combination with bibliographic means like citations. In this work we will present an approach that combines a cognitive information retrieval framework based on the principle of polyrepresentation with document clustering to enable the user to explore a collection more interactively than by just examining a ranked result list. Our approach uses information need representations as well as different document representations including citations. To evaluate our ideas we employ a simulated user strategy utilising a cluster ranking approach. We report on the possible effectiveness of our approach and on several strategies how users can achieve a higher search effectiveness through cluster browsing. Our results confirm that our proposed polyrepresentative cluster browsing strategy can in principle significantly improve the search effectiveness. However, further evaluations including a more refined user simulation are needed.
    • Exploiting information needs and bibliographics for polyrepresentative document clustering

      Abbasi, Muhammad Kamran; Frommholz, Ingo; University of Bedfordshire (CEUR-WS, 2014-04)
      In this paper we explore the potential of combining the principle of polyrepresentation with document clustering. Our idea is discussed and evaluated for polyrepresentation of information needs as wells as for document-based polyrepresentation where bibliographic information is used as representation. The main idea is to present the user with the highly ranked polyrepresentative clusters to support the search process. Our evaluation suggests that our approach is capable of increasing retrieval performance, but performance varies for queries with a high or low number of relevant documents.
    • On clustering and polyrepresentation

      Frommholz, Ingo; Abbasi, Muhammad Kamran; University of Bedfordshire (Springer Verlag, 2014-04)
      Polyrepresentation is one of the most prominent principles in a cognitive approach to interactive information seeking and retrieval. When it comes to interactive retrieval, clustering is another method for accessing information. While polyrepresentation has been explored and validated in a scenario where a system returns a ranking of documents, so far there are no insights if and how polyrepresentation and clustering can be combined. In this paper we discuss how both are related and present an approach to integrate polyrepresentation into clustering. We further report some initial evaluation results.