The dominant method for evaluating search engines is the Cranfield paradigm, but the existing metrics do not consider some modern search engines features, such as document snippets...
Search engines present fix-length passages from documents ranked by relevance against the query. In this paper, we present and compare novel, language-model based methods for extr...
Determining the similarity of short text snippets, such as search queries, works poorly with traditional document similarity measures (e.g., cosine), since there are often few, if...
Users of Web search engines are often forced to sift through the long ordered list of document “snippets” returned by the engines. The IR community has explored document cluste...
Throughout many of the different types of Web searches people perform, the primary tasks are to first craft a query that effectively captures their information needs, and then eva...