Zhenzhen Kou
ABSTRACT
There have been great needs to develop effective methods for combining multiple rankings from heterogeneous domains into one single rank list arising from many recent web search applications, such as integrating web search results from multiple engines, facets, or verticals. We define this problem as Learning to blend rankings from multiple domains. We propose a class of learning-to-blend methods that learn a monotonically increasing transformation for each ranking so that the rank order in each domain is preserved and the transformed values are comparable across multiple rankings. The transformation learning can be tackled by solving a quadratic programming problem. The novel machine learning method for blending multiple ranking lists is evaluated with queries sampled from a commercial search engine and a promising improvement of Discounted Cumulative Gain has been observed.
- Chris Burges, Tal Shaked, Erin Renshaw, Ari Lazier, Matt Deeds, Nicole Hamilton, and Greg Hullender. Learning to rank using gradient descent. In ICML '05: Proceedings of the 22nd international conference on Machine learning, pages 89--96, 2005. Google Scholar
Digital Library
- Zhe Cao, Tao Qin, Tie-Yan Liu, Ming-Feng Tsai, and Hang Li. Learning to rank: from pairwise approach to listwise approach. In ICML, pages 129--136, 2007. Google ScholarDigital Library
- C. Cortes, M. Mohri, and A. Rastogi. Magnitude-preserving ranking algorithms. In Proceedings of the 24th ICML, 2007. Google ScholarDigital Library
- Cynthia Dwork, Ravi Kumar, Moni Naor, D. Sivakuma. Rank Aggregation Methods for the Web, In the Proceedings of the 10th international conference on World Wide Web, pages 613--622, 2001. Google ScholarDigital Library
- Y. Freund, R. Iyer, R. Schapire, and Y. Singer. An efficient boosting algorithm for combining preferences. In Proceedings of the Fifteenth International Conference on Machine Learning, 1998. Google ScholarDigital Library
- J. Guiver and E. Snelson. Learning to rank with SoftRank and Gaussian processes. In Proceedings of the 31st annual international ACM SIGIR conference on Research and development in information retrieval, 2008. Google ScholarDigital Library
- K. Jarvelin and J. Kekalainen. Cumulated gain-based evaluation of IR techniques. ACM Transactions on Information Systems, 20:422--446, 2002. Google ScholarDigital Library
- T. Joachims. Optimizing search engines using clickthrough data. In Proceedings of ACM SIGKDD, 2002. Google ScholarDigital Library
- Yu-Ting Liu, Tie-Yan Liu, Tao Qin, Zhi-Ming Ma, Hang Li. Supervised rank aggregation, In Proceedings of the 16th international conference on World Wide Web, pages: 481--490, 2007. Google ScholarDigital Library
- Taesup Moon, Alex Smola, Yi Chang and Zhaohui Zheng. IntervalRank - Isotonic Regression with Listwise and Pairwise Constraints. In WSDM, pages 151--160, 2010 Google ScholarDigital Library
- J. Xu and H. Li. Adarank: a boosting algorithm for information retrieval. In Proceedings of the 30th ACM SIGIR, 2007. Google ScholarDigital Library
- Zhaohui Zheng, Hongyuan Zha, and Gordon Sun. Query-level learning to rank using isotonic regression. In the 46th Annual Allerton Conference on Communication, Control and Computing, 2008.Google ScholarCross Ref
Index Terms
- Learning to blend rankings: a monotonic transformation to blend rankings from heterogeneous domains
Recommendations
What's Going on in Search Engine Rankings?
AINAW '08: Proceedings of the 22nd International Conference on Advanced Information Networking and Applications - WorkshopsMany people use search engines every day to retrieve documents from the Web. Although the social influence of search engine rankings has become significant, ranking algorithms are not disclosed. In this paper, we have investigated three major search ...
Comparing rankings of search results on the Web
Special issue: InfometricsThe Web has become an information source for professional data gathering. Because of the vast amounts of information on almost all topics, one cannot systematically go over the whole set of results, and therefore must rely on the ordering of the results ...
Methods for comparing rankings of search engine results
Web dynamicsIn this paper we present a number of measures that compare rankings of search engine results. We apply these measures to five queries that were monitored daily for two periods of 14 or 21 days each. Rankings of the different search engines (Google, ...
Reviews
George Popescu
The problem of aggregating rankings from heterogeneous domains into one, all-comprising rank is very difficult in its formulation. It takes into account ranking relevance based on the domains where rankings are aggregated: books, perfumes, music, movies, and so on. Thus, the solution of having one universal function is limited, and more sophisticated methods that blend documents and domains together are more appropriate. To establish ground truth, one needs to label the results generated by a search engine with some standard categories, ranging from "perfect" to "bad." The monotonic increasing transformation is the key concept, and it acts on the set of training data to generate a combined ranking score. By solving a quadratic linear programming equation, one determines the ranking coefficients. The data experiments, based on queries from Yahoo! Answers, show the results and best-case scenarios for blended learning in which a discounted cumulative gain (DCG) of ten is the upper bound. The demand for vertical searches over the Web describes the need for more efficient methods to aggregate page ranks from multiple domains and display the most relevant ones. The algorithm presented in the paper shows how blended learning can be an adaptive process that helps to determine the optimal combined list of ranked items. Online Computing Reviews Service
Access critical reviews of Computing literature here
Become a reviewer for Computing Reviews.
Comments