Using Access Data for Paper Recommendations on ArXiv.org
Stefan Pohl

TL;DR
This paper explores using access log data from arXiv.org to develop a paper recommendation system, comparing it with content and citation-based methods, aiming to provide immediate, user-behavior-driven suggestions for scientists.
Contribution
It introduces a novel approach of leveraging access logs for paper recommendations, demonstrating its effectiveness as a real-time alternative to traditional citation-based methods.
Findings
Access logs can effectively identify related papers.
User behavior-based recommendations perform comparably to citation-based methods.
An online recommendation system was successfully implemented.
Abstract
This thesis investigates in the use of access log data as a source of information for identifying related scientific papers. This is done for arXiv.org, the authority for publication of e-prints in several fields of physics. Compared to citation information, access logs have the advantage of being immediately available, without manual or automatic extraction of the citation graph. Because of that, a main focus is on the question, how far user behavior can serve as a replacement for explicit meta-data, which potentially might be expensive or completely unavailable. Therefore, we compare access, content, and citation-based measures of relatedness on different recommendation tasks. As a final result, an online recommendation system has been built that can help scientists to find further relevant literature, without having to search for them actively.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsRecommender Systems and Techniques · Data Mining Algorithms and Applications · Web Data Mining and Analysis
