Are click-through data adequate for learning web search rankings?

ABSTRACT

Learning-to-rank algorithms, which can automatically adapt ranking functions in web search, require a large volume of training data. A traditional way of generating training examples is to employ human experts to judge the relevance of documents. Unfortunately, it is difficult, time-consuming and costly. In this paper, we study the problem of exploiting click-through data for learning web search rankings that can be collected at much lower cost. We extract pairwise relevance preferences from a large-scale aggregated click-through dataset, compare these preferences with explicit human judgments, and use them as training examples to learn ranking functions. We find click-through data are useful and effective in learning ranking functions. A straightforward use of aggregated click-through data can outperform human judgments. We demonstrate that the strategies are only slightly affected by fraudulent clicks. We also reveal that the pairs which are very reliable, e.g., the pairs consisting of documents with large click frequency differences, are not sufficient for learning.

REFERENCES

• 1: Eugene Agichtein , Eric Brill , Susan Dumais, Improving web search ranking by incorporating user behavior information, Proceedings of the 29th annual international ACM SIGIR conference on Research and development in information retrieval, August 06-11, 2006, Seattle, Washington, USA [doi>10.1145/1148170.1148177]
• 2: Eugene Agichtein , Eric Brill , Susan Dumais , Robert Ragno, Learning user interaction models for predicting web search result preferences, Proceedings of the 29th annual international ACM SIGIR conference on Research and development in information retrieval, August 06-11, 2006, Seattle, Washington, USA [doi>10.1145/1148170.1148175]
• 3: Andrei Broder, A taxonomy of web search, ACM SIGIR Forum, v.36 n.2, Fall 2002 [doi>10.1145/792550.792552]
• 4: Chris Burges , Tal Shaked , Erin Renshaw , Ari Lazier , Matt Deeds , Nicole Hamilton , Greg Hullender, Learning to rank using gradient descent, Proceedings of the 22nd international conference on Machine learning, p.89-96, August 07-11, 2005, Bonn, Germany [doi>10.1145/1102351.1102363]
• 5: C.J.C. Burges, R. Ragno, and Q.V. Le. Learning to rank with nonsmooth cost functions. In Advances in Neural Information Processing Systems 18, pages 395--402, Cambridge, MA, 2006. MIT Press.
• 6: Yunbo Cao , Jun Xu , Tie-Yan Liu , Hang Li , Yalou Huang , Hsiao-Wuen Hon, Adapting ranking SVM to document retrieval, Proceedings of the 29th annual international ACM SIGIR conference on Research and development in information retrieval, August 06-11, 2006, Seattle, Washington, USA [doi>10.1145/1148170.1148205]
• 7: Zhicheng Dou , Ruihua Song , Ji-Rong Wen, A large-scale evaluation and analysis of personalized search strategies, Proceedings of the 16th international conference on World Wide Web, May 08-12, 2007, Banff, Alberta, Canada [doi>10.1145/1242572.1242651]
• 8: Steve Fox , Kuldeep Karnawat , Mark Mydland , Susan Dumais , Thomas White, Evaluating implicit measures to improve web search, ACM Transactions on Information Systems (TOIS), v.23 n.2, p.147-168, April 2005 [doi>10.1145/1059981.1059982]
• 9: Yoav Freund , Raj Iyer , Robert E. Schapire , Yoram Singer, An efficient boosting algorithm for combining preferences, The Journal of Machine Learning Research, 4, p.933-969, 12/1/2003
• 10: Kalervo Järvelin , Jaana Kekäläinen, IR evaluation methods for retrieving highly relevant documents, Proceedings of the 23rd annual international ACM SIGIR conference on Research and development in information retrieval, p.41-48, July 24-28, 2000, Athens, Greece [doi>10.1145/345508.345545]
• 11: Thorsten Joachims, Optimizing search engines using clickthrough data, Proceedings of the eighth ACM SIGKDD international conference on Knowledge discovery and data mining, July 23-26, 2002, Edmonton, Alberta, Canada [doi>10.1145/775047.775067]
• 12: Thorsten Joachims , Laura Granka , Bing Pan , Helene Hembrooke , Geri Gay, Accurately interpreting clickthrough data as implicit feedback, Proceedings of the 28th annual international ACM SIGIR conference on Research and development in information retrieval, August 15-19, 2005, Salvador, Brazil [doi>10.1145/1076034.1076063]
• 13: Thorsten Joachims , Laura Granka , Bing Pan , Helene Hembrooke , Filip Radlinski , Geri Gay, Evaluating the accuracy of implicit feedback from clicks and query reformulations in Web search, ACM Transactions on Information Systems (TOIS), v.25 n.2, p.7-es, April 2007 [doi>10.1145/1229179.1229181]
• 14: Diane Kelly , Jaime Teevan, Implicit feedback for inferring user preference: a bibliography, ACM SIGIR Forum, v.37 n.2, p.18-28, Fall 2003 [doi>10.1145/959258.959260]
• 15: M. Kendall and B.B. Smith. Randomness and random sampling numbers. Journal of the Royal Statistical Society, 101(1):147--166, 1938.
• 16: T.-Y. Liu, T. Qin, J. Xu, W. Xiong, and H. Li. Letor: Benchmark dataset for research on learning to rank for information retrieval. In LR4IR 2007 in conjunction with SIGIR 2007, 2007.
• 17: {17} F. Radlinski and T. Joachims. Evaluating the robustness of learning from implicit feedback. In Proceedings of the 22nd ICML Workshop on Learning in Web Search, 2005.
• 18: Filip Radlinski , Thorsten Joachims, Query chains: learning to rank from implicit feedback, Proceedings of the eleventh ACM SIGKDD international conference on Knowledge discovery in data mining, August 21-24, 2005, Chicago, Illinois, USA [doi>10.1145/1081870.1081899]
• 19: F. Radlinski and T. Joachims. Minimally invasive randomization for collecting unbiased preferences from clickthrough logs. In Proceedings of the 21st National Conference on Artificial Intelligence (AAAI), 2006.
• 20: Zhaohui Zheng , Keke Chen , Gordon Sun , Hongyuan Zha, A regression framework for learning ranking functions using relative relevance judgments, Proceedings of the 30th annual international ACM SIGIR conference on Research and development in information retrieval, July 23-27, 2007, Amsterdam, The Netherlands [doi>10.1145/1277741.1277792]

INDEX TERMS

Primary Classification:
  H. Information Systems
  H.3 INFORMATION STORAGE AND RETRIEVAL
      H.3.3 Information Search and Retrieval
          Subjects: Search process

Additional Classification:
  H. Information Systems
  H.3 INFORMATION STORAGE AND RETRIEVAL
      H.3.3 Information Search and Retrieval
          Subjects: Relevance feedback

General Terms:
Algorithms, Experimentation, Measurement, Performance

Keywords:
click-through data, implicit feedback, learning to rank, relevance judgments, web search rankings