Rhetorical Classification of Anchor Text for Citation Recommendation

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