Social summarization in collaborative web search Oisı ´n Boydell, Barry Smyth * CLARITY: Centre for Sensor Web Technologies, School of Computer Science and Informatics, University College Dublin Belfield, Dublin 4, Ireland article info Article history: Received 15 May 2009 Received in revised form 1 September 2009 Accepted 30 October 2009 Available online 22 December 2009 Keywords: Summarization Personalization Web search abstract A critical challenge for Web search engines concerns how they present relevant results to searchers. The traditional approach is to produce a ranked list of results with title and sum- mary (snippet) information, and these snippets are usually chosen based on the current query. Snippets play a vital sensemaking role, helping searchers to efficiently make sense of a collection of search results, as well as determine the likely relevance of individual results. Recently researchers have begun to explore how snippets might also be adapted based on searcher preferences as a way to better highlight relevant results to the searcher. In this paper we focus on the role of snippets in collaborative web search and describe a technique for summarizing search results that harnesses the collaborative search behav- iour of communities of like-minded searchers to produce snippets that are more focused on the preferences of the searchers. We go on to show how this so-called social summari- zation technique can generate summaries that are significantly better adapted to searcher preferences and describe a novel personalized search interface that combines result recom- mendation with social summarization. Ó 2009 Elsevier Ltd. All rights reserved. 1. Introduction From a Web search standpoint, the success of a particular result-list depends on a number of factors. Obviously the pages that are retrieved as results are important; missing relevant result pages or including too many irrelevant result pages will significantly compromise the quality of the result-list. In addition, the ability to rank results according to their likely rele- vance to the query is also critically important and it is well known that the majority of user attention tends to be focused on the top ranking results. Finally, results should be presented in a way that highlights their likely relevance, not just to the query, but to the individual searcher. By convention, today’s search engines present results as a combination of page title, page URL, and result snippet. In this paper we are especially interested in result snippets—those short extracts of page content that serve to summarize a particular result—and the way that they are generated. In the past researchers have attempted to improve Web search by concentrating on the selection and ranking of search results. For example, many researchers have called for a more personalized approach to Web search, one which takes advan- tage of the learned preferences of the individual searcher (Dou, Song, & Wen, 2007) or a community of searchers, so as to recommend a ranked list of results that better reflect these interests. This research shares many aspects in common with traditional recommender systems research as it involves the recommendation of items (search results) on the basis of some learned user (or community) preferences. Recently, recommender systems research has begun to look at how recommenda- tions can be explained or justified to users, to help users better understand the reason behind a recommendation, and ulti- mately improve the perceived quality of the recommendations that are made (McSherry, 2005; Pu & Chen, 2007). In this 0306-4573/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.ipm.2009.10.011 * Corresponding author. Tel.: +353 1 7162473. E-mail address: barry.smyth@ucd.ie (B. Smyth). Information Processing and Management 46 (2010) 782–798 Contents lists available at ScienceDirect Information Processing and Management journal homepage: www.elsevier.com/locate/infoproman