2010 ContextAwareCitationRecommendation

From GM-RKB
Jump to navigation Jump to search

Subject Headings: Citation Recommendation Task, Recommendation Algorithm.

Notes

Quotes

Author Keywords

Bibliometrics, Context, Gleason's Theorem, Recommender Systems.

Abstract

When you write papers, how many times do you want to make some citations at a place but you are not sure which papers to cite? Do you wish to have a recommendation system which can recommend a small number of good candidates for every place that you want to make some citations? In this paper, we present our initiative of building a context-aware citation recommendation system. High quality citation recommendation is challenging: not only should the citations recommended be relevant to the paper under composition, but also should match the local contexts of the places citations are made. Moreover, it is far from trivial to model how the topic of the whole paper and the contexts of the citation places should affect the selection and ranking of citations. To tackle the problem, we develop a context-aware approach. The core idea is to design a novel non-parametric probabilistic model which can measure the context-based relevance between a citation context and a document. Our approach can recommend citations for a context effectively. Moreover, it can recommend a set of citations for a paper with high quality. We implement a prototype system in CiteSeerX. An extensive empirical evaluation in the CiteSeerX digital library against many baselines demonstrates the effectiveness and the scalability of our approach.

1. Introduction

When you write papers, how many times do you want to make some citations at a place but you are not sure which papers to cite? For example, the left part of Figure 1 shows a segment of a query manuscript containing some citation placeholders (placeholders for short) marked as "[?]", where citations should be added. In order to fill in those citation placeholders, one needs to search the relevant literature and find a small number of proper citations. Searching for proper citations is often a labor-intensive task in research paper composition, particularly for junior researchers who are not familiar with the very extensive literature. Moreover, the volume of research undertaken and information available make citation search hard even for senior researchers. Do you wish to have a recommendation system which can recommend a small number of good candidates for every place that you want to make some citations? High quality citation recommendation is a challenging problem for many reasons.

For each citation placeholder, we can collect the words surrounding as the context of the placeholder. One may think we can use some keywords in the context of a placeholder to search a literature search engine like Google Scholar or CiteSeerX to obtain a list of documents as the candidates for citations. However, such a method, based on keyword matching, is often far from satisfactory. For example, using query "frequent itemset mining" one may want to search for the first paper proposing the concept of frequent itemset mining, e.g. [1]. However, Google Scholar returns a paper about frequent closed itemset mining published in 2000 as the first result, and a paper on privacy preserving frequent itemset mining as the second choice. [1] does not appear in the first page of the results. CiteSeerX also lists a paper on privacy preserving frequent itemset mining as the first result. CiteSeerX fails to return [1] on the first page, either. One may wonder, as we can model citations as links from citing documents to cited ones, can we use graph-based link prediction techniques to recommend citations? Graph-based link prediction techniques often require a user to provide sample citations for each placeholder, and thus shifts much of the burden to the user. And, graph-based link prediction methods may encounter difficulties to make proper citations across multiple communities because a community may not be aware of the related work in some other community.


,

 AuthorvolumeDate ValuetitletypejournaltitleUrldoinoteyear
2010 ContextAwareCitationRecommendationJian Pei
Qi He
Daniel Kifer
Prasenjit Mitra
C. Lee Giles
Context-aware Citation RecommendationProceedings of the 19th International World Wide Web Conferencehttp://www.cs.sfu.ca/~jpei/publications/CitationRecommendation WWW2010.pdf2010