Presented at: 17th International World Wide Web Conference (WWW2008)
by Ariel Fuxman, Panayiotis Tsaparas, Kannan Achan, Rakesh Agrawal
Webpage: http://doi.acm.org/10.1145/1367497.1367506Webpage: http://www2008.org/papers/pdf/p61-fuxmanA.pdf
In the sponsored search model, search engines are paid by businesses that are interested in displaying ads for their site alongside the search results. Businesses bid for different keywords, and their ad is displayed when the keyword is queried to the search engine. An important problem in this process is \emph{keyword generation}: given a business that is interested in being advertised with respect to a specific concept, suggest keywords that are related to this concept. We address this problem by making use of the query logs of the search engine. We identify queries related to a concept by exploiting the associations between queries and URLs as they are captured by the user's clicks. These queries form good keyword suggestions since they capture the ``wisdom of the crowd'' as to what is related to a site, and they can be directly monetized. We formulate the problem as a semi-supervised learning problem, and we propose two different algorithms within the Markov Random Field model. We perform extensive experiments with real query logs and we demonstrate that our algorithms can scale for large query logs, and produce meaningful results.
Keywords: click logs, keyword generation, markov random fields, random walks, World Wide Web
Resource URI on the dog food server: http://data.semanticweb.org/conference/www/2008/paper/956
Explore this resource elsewhere: