Topical search engines and query context models
AssigneeMicrosoft Technology Licensing, LLC
InventorsRaman Chandrasekar, Parikshit Sondhi, Robert Rounthwaite
Life of this patent
Estimating the term…
Abstract
Topical search engines can add contextual keywords to an input query to bias results toward a particular topic or domain. In one instance, query context models can be constructed to facilitate topical search. Upon receipt of one or more topic-relevant sites, a plurality of topical queries can identified automatically. Contextual keywords can be identified with respect to the plurality of topical queries as a function of lexical generality, among other things. Subsequently, a query context model, comprising the identified topical queries and related contextual keywords, can be employed to restrict query results to a particular topic.
Claims20
A method of facilitating search, comprising:
- employing at least one processor configured to execute computer-executable instructions stored in memory to perform the following acts:
- identifying at least one contextual keyword from a structure based on a received query, the structure comprises one or more contextual keywords for a number of queries associated with a specific topic based on at least one topic-relevant seed uniform resource locator, one or more queries associated with the at least one seed uniform resource locator, extracted from a record of user queries and selected uniform resource locators for each user query, one or more associated uniform resources locators derived from the one or more associated queries, and one or more additional queries derived from the one or more associated uniform resource locators, wherein the additional queries are not directly associated with the at least one seed uniform resource locator in the record;
- and
- adding the at least one contextual keyword to the received query restricting the query to the specific topic.
A search system, comprising:
- a processor coupled to a memory, the processor configured to execute the following computer-executable component stored in the memory:
- a first component configured to automatically identify a set of queries related to a particular topic as a function of a one or more topic-relevant seed uniform resource locators (URLs), a record of previously submitted queries and selected URLs, and one or more additional URLs derived from one or more queries associated with the one or more seed URLs, wherein one or more additional queries of the set of queries derived from the additional URLs are not directly associated with the one or more seed URLs in the record.
A computer-readable storage medium having instructions stored thereon that enable at least one processor to perform a method upon execution of the instructions, the method comprising:
- identifying one or more queries based on one or more topic-relevant seed uniform resource locators (URLs), a record of previously submitted queries and selected URLs, and one or more additional URLs derived from one or more queries associated with the one or more seed URLs, wherein one or more additional queries of the one of the one or more queries identified from the additional URLs are not directly associated with the one or more seed URLs in the record;
- determining a co-occurrence score for one or more keywords comprising the one or more queries;
- determining a lexical generality score for the one or more keywords;
- and
- saving the one or more keywords, and co-occurrence and lexical generality scores to a persistent store.
Dependencies are inferred from the claim text.