WEB-RELEVANCE BASED QUERY CLASSIFICATION

Methods, systems, algorithms, and media are provided for query classification based upon web relevance. In embodiments, a knowledge domain is selected. A knowledge domain includes a set of seed-web domains which are classified to belong to the knowledge domain. For each seed-web domain, a filtered web search of a given query is implemented utilizing a filter associated with a particular seed-web domain to obtain a total number of filtered results for the seed-web domain. For each seed-web domain, a total number of non-filtered results are referenced. A percentage of coverage for each seed-web domain is calculated using the total number of filtered results for the seed-web domain and the total number of non-filtered results for the seed-web domain. An average percentage of coverage for the knowledge domain is calculated using the percentage of coverage for each seed-web domain.