p116-mcnee - On the Recommending of Citations for Research...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon
On the Recommending of Citations for Research Papers Sean M. McNee, Istvan Albert, Dan Cosley, Prateep Gopalkrishnan, Shyong K. Lam, Al Mamunur Rashid, Joseph A. Konstan, John Riedl GroupLens Research Project Department of Computer Science and Engineering University of Minnesota Minneapolis, MN 55455 USA {mcnee, ialbert, cosley, prateep, lam, arashid, konstan, riedl}@cs.umn.edu ABSTRACT Collaborative filtering has proven to be valuable for recommending items in many different domains. In this paper, we explore the use of collaborative filtering to recommend research papers, using the citation web between papers to create the ratings matrix. Specifically, we tested the ability of collaborative filtering to recommend citations that would be suitable additional references for a target research paper. We investigated six algorithms for selecting citations, evaluating them through offline experiments against a database of over 186,000 research papers contained in ResearchIndex. We also performed an online experiment with over 120 users to gauge user opinion of the effectiveness of the algorithms and of the utility of such recommendations for common research tasks. We found large differences in the accuracy of the algorithms in the offline experiment, especially when balanced for coverage. In the online experiment, users felt they received quality recommendations, and were enthusiastic about the idea of receiving recommendations in this domain. Keywords Collaborative Filtering, Recommender Systems, Citation Graphs, Social Networks, Digital Libraries, ResearchIndex INTRODUCTION People face the problem of information overload every day, a problem that is only getting worse. As more and more people publish information on the World Wide Web, it becomes increasingly difficult to find needed information quickly. By recommending items to users based on previously expressed user preferences, recommender systems help users navigate and control complex information spaces. The MovieLens recommender (www.movielens.org), for example, helps users of the system find movies to watch solely based on their opinions of films. Collaborative Filtering (CF) is a widely used technique in recommender systems. CF works by matching users in a system based on the similarity of each user’s past preferences. Each user has a ‘neighborhood’ of other users with similar opinions about items in the system. This neighborhood can be used to generate recommendations by suggesting items to the user that he has not viewed but that his neighbors have viewed and rated highly. Collaborative filtering has several limitations. One of the most important is the startup problem [15]. When a CF system is first created, there are many items in the system, few users in the system, and no ratings. Without ratings, the system cannot generate recommendations and users see no benefit. Without users, there is no way for new ratings to be entered into the system. When applying CF to a domain, it is valuable to seek preexisting data that can be used to seed such a database of ratings.
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 2
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 02/24/2010 for the course COMM 4400 at Cornell University (Engineering School).

Page1 / 10

p116-mcnee - On the Recommending of Citations for Research...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online