Arbitration of specialized content using search results

• US 20070078822A1
• Filed: 09/30/2005
• Published: 04/05/2007
• Est. Priority Date: 09/30/2005
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Cucerzan, Silviu-Petru, Kern, Randall, Buckley, Jamie, Brill, Eric, Baldwin, Laura
• Granted Patent: US 7,895,193 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques

Abstract

The claimed subject matter provides systems and methods that facilitate providing specialized content associated with web results in response to a user query to provide queried information directly to a user without requiring the user to select a link and/or scroll through information on a web page associated therewith. Reference language models can be generated from multiple queries and can be compared to a query language model generated from web results returned in response to a user query. Based on such comparison, a determination can be made regarding whether a particular portion of specialized content corresponding to some particular query keywords is relevant enough to warrant direct presentation to the user.

Claims

1. A computer implemented system that facilitates providing specialized content to a user in response to a user query, comprising:

an interface component by which a user enters a query and receives returned data; and

an arbitration component that returns specialized content upon a determination that the specialized content is relevant to the user'"'"'s query.

2. The system of claim 1, in which the arbitration component uses a query processing component that identifies whether a query contains terms relevant to at least one of a plurality of specialized content topics.

3. The system of claim 1, in which the arbitration component comprises:

a language model generator that builds a query language model based on results returned from a search engine in response to the user query;

at least one reference language model for each specialized content topic; and

a comparison component that evaluates the similarity between at least one reference language model and the query language model.

4. The system of claim 3, further comprising a reference language model generator that utilizes topical terms to create at least one reference language model for each of a plurality of searchable topics.

5. The system of claim 3, further comprising a threshold component that determines whether the similarity between the query language model and at least one of the reference language models is above a predetermined threshold relevance level for the corresponding topic in order to present specialized content for the user'"'"'s query.

6. The system of claim 1, further comprising a personalization component that adjusts at least one relevance threshold for specialized content results based at least in part on user preferences.

7. The system of claim 6, further comprising an implicit component that adjusts the relevance threshold level based on implicit user preferences.

8. The system of claim 7, the implicit user preferences are determined from at least one of user query history and a user query log.

9. The system of claim 6, further comprising an explicit component that adjusts the relevance threshold based on explicit user preferences related to at least one preferred specialized content topic.

10. A computer-implemented method that facilitates providing specialized content to a user, comprising:

identifying at least one term in a user query as being related to at least one specialized content topic;

generating a query language model upon receipt of returned search results for a user query;

comparing the query language model to at least one reference language model corresponding to a specialized content topic related to terms in the query;

determining whether the query language model and the at least one reference language model display similarity equal to or above a relevance threshold value; and

displaying specialized content if the relevance threshold value is equaled or exceeded.

11. The method of claim 10, further comprising displaying returned search results without specialized content if the relevance threshold value is not met or exceeded.

12. The method of claim 10, further comprising generating at least one reference model for each of a plurality of query topics.

13. The method of claim 12, generating the reference model comprises:

analyzing returned results for a plurality of topic-specific queries;

analyzing returned results for a plurality of general queries; and

weighting words that occur with a higher frequency in the topic-specific queries than in the general queries.

14. The method of claim 10, further comprising arbitrating between topics for which the relevance threshold value is equaled or exceeded to display specialized content associated with at least one of these topics.

15. The method of claim 14, further comprising arbitrating between specialized content snippets within a topic associated with topical terms in the query by evaluating an aggregate weight for each specialized content snippet.

16. The method of claim 10, further comprising implicitly adjusting the relevance threshold for a given query topic based at least in part on user query history.

17. The method of claim 10, further comprising explicitly adjusting the relevance threshold for a given query topic based at least in part on user-indicated preferences.

18. A computer-implemented system that facilitates providing specialized content to a user in response to a query, comprising:

means for identifying terms in a user query as being related to a particular topic;

means for comparing properties in user query results to weighted reference properties;

means for determining whether specialized content associated with topical terms in the user query is above a relevance threshold; and

means for displaying specialized content to a user upon a determination that the specialized content is above the relevance threshold.

19. The system of claim 18, further comprising means for adjusting the relevance threshold in response to user preferences.

20. The system of claim 18, further comprising means for generating at least one reference language model for each of a plurality of query topics and weighting reference properties associated therewith.