Abstract:
The immense proliferation of research papers in journals and conferences poses
challenges for researchers wanting to access relevant scholarly papers. Recommender
systems o er a solution to this research problem by ltering all of the
available information and delivering what is most relevant to the user.
Several approaches have been proposed for research paper recommendation, variously
based on metadata, content, citation analysis, collaborative ltering, etc.
Approaches predicated on citation analysis, including co-citation analysis and bibliographic
coupling, have proven to be signi cant. Co-citation has been analyzed
at content level and the use of citation proximity analysis has shown signi cant
improvement in accuracy. However, co-citation presents the relationship between
two papers based on their having been mutually cited by other papers, without
considering the contents of the citing papers. Bibliographic coupling, on the other
hand, considers two papers as relevant if they share common references, but traditionally
does not consider the citing patterns of common references in di erent
logical parts of the citing papers.
The improvement found in cases of co-citation when combined with content analysis,
motivated us to analyze the impact of using proximity analysis of in-text
citations in cases of bibliographic coupling. Therefore, in this research, three different
approaches were proposed that extended bibliographic coupling by exploiting
the proximity of in-text citations of bibliographically coupled articles. These
approaches are: (1) DBSCAN-based bibliographic coupling, (2) centiles-based bibliographic
coupling and (3) section-based bibliographic coupling. Comprehensive
experiments utilizing both user study and automated evaluations were conducted
to evaluate the proposed approaches. The results showed signi cant improvement
over traditional bibliographic coupling and content-based research paper recommendation
