Bandits and Browsing                              Effective Collection Size as Way of Quantifying Search EfficiencyHarriet...
Upcoming SlideShare
Loading in …5

Bandits and Browsing: Effective Collection Size as Way of Quantifying Search Efficiency


Published on

This poster presents our preliminary research on how information can be extracted from user browsing behavior to identify understudied works that are relevant but have too few viewers. We investigate how to apply two types of analysis—a formula called Effective Collection Size and ‘multi-armed bandit’ analysis—to extracted user data to develop alternative methods of retrieving materials from collection that are collated by richer factors of relevancy. We anticipate that these analyses will enable the development of an information retrieval system that presents a broad range of content in a user’s search results.

Published in: Education, Technology
  • Be the first to comment

  • Be the first to like this

Bandits and Browsing: Effective Collection Size as Way of Quantifying Search Efficiency

  1. 1. Bandits and Browsing Effective Collection Size as Way of Quantifying Search EfficiencyHarriett E. Green, Kirk Hess, and Richard D. Hislop  University of Illinois at Urbana-Champaign   rhislop2@illinois.eduEFFECTIVE COLLECTION ANALYSIS AND INITIAL RESULTSSIZE • Ran statistical analysis on the English collection. • Found books and topics that are of unusually highEffective Collection Size quantifies how use and quantified statistically.efficiently a library uses its collection. It • Identified improbably understudied items.focuses on highlighting understudied • Found topics of interest for digital collectionworks and aims to prevent the omission development.of useful materials in a collection. WHY? NEXT STEPS • Analyze the broader University of Illinois catalog. Biases in traditional search • Incorporate analysis into Illinois Harvest digital algorithms send most users to the library search results. same high-ranking materials. Digital Circulation of all titles with threshold • Produce a set of tools to help highlight libraries can adapt to user behavior, of 100 checkouts understudied materials during reference and identify useful material and send digitization projects. users to relevant but understudied • Use results to quantify increases in efficiency of sources. collection use. OUR PROJECT SELECT REFERENCES Zhou, T., Kuscsik, Z., Liu, J.G., Medo, M., Wakeling, J.R., & Zhang, Y.C. • Prototype data: University of (2010). Solving the apparent diversity-accuracy dilemma of Illinois Library catalog circulation recommender systems. Proceedings of the National Academy of Sciences of the United States of America, 107, 4511-4515. statistics • Use our physical catalog to learn Li, L., Chu, W., Langford, J., & Schapire, R. E. (2010). A contextual- about collection use bandit approach to personalized news article recommendation. Proceedings of the Nineteenth International Conference on World Wide • Apply this to improve search and Web, 661-670. Doi: 10.1145/1772690.1772758 recommendations in digital collections Circulation of all titles with more than Xie, I. & Cool, C. (2009). Understanding help seeking within the 100 checkouts context of searching digital libraries. Journal of the American Society for Information Science and Technology, 60, 477--494.  Twitter: @greenharr 2011 DLF Forum October 31-November 1, 2011