Rhetorical Classification of Anchor Text for Citation Recommendation

Daniel Duma, Maria Liakata, Amanda Clare, James Ravenscroft, Ewan Klein

Research output: Contribution to journalArticlepeer-review

265 Downloads (Pure)

Abstract

Wouldn’t it be helpful if your text editor automatically suggested papers that are contextually relevant to your work? We concern ourselves with this task: we desire to recommend contextually relevant citations to the author of a paper.
A number of rhetorical annotation schemes for academic articles have been developed over the years, and it has often been suggested that they could find application in Information Retrieval scenarios such as this one. In this
paper we investigate the usefulness for this task of CoreSC, a sentence-based, functional, scientific discourse annotation scheme (e.g. Hypothesis, Method, Result, etc.). We specifi-cally apply this to anchor text, that is, the text surrounding a citation, which is an important source of data for building
document representations. By annotating each sentence in every document with CoreSC and indexing them separately by sentence class, we aim to build a more useful vector-space representation of documents in our collection. Our results
show consistent links between types of citing sentences and types of cited sentences in anchor text, which we argue can indeed be exploited to increase the relevance of recommendations
Original languageEnglish
JournalD-Lib Magazine
Volume22
Issue number9/10
DOIs
Publication statusPublished - 15 Sept 2016

Keywords

  • core scientific concepts
  • CoreSC
  • context based
  • citation recommendation
  • anchor text
  • incoming link contexts

Fingerprint

Dive into the research topics of 'Rhetorical Classification of Anchor Text for Citation Recommendation'. Together they form a unique fingerprint.

Cite this