System for determining local intent in a search query

• US 7,788,252 B2
• Filed: 03/28/2007
• Issued: 08/31/2010
• Est. Priority Date: 03/28/2007
• Status: Expired due to Fees

• Current Assignee: R2 Solutions LLC (Acacia Research Corporation)
• Original Assignee: Yahoo! Inc. (Apollo Global Management, Inc.)
• Inventors: Stipp, Eugene, Demir, Ramazan, Delli Santi, Jim W.
• Primary Examiner(s): LeRoux; Etienne P
• Assistant Examiner(s): Le; Hung D
• Application Number:
  US11/729,096
• Publication Number:
  US 20080243611A1
• Time in Patent Office: 1,252 Days
• Field of Search: None
• US Class Current: 707/708
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0267 Wireless devices

Abstract

A system and method are disclosed for determining local intent. Local intent may reflect whether a search query should receive results and advertisements that are geographically specific. The local intent may be determined using probabilistic models that analyze historical searches to determine which search terms tend to have local intent.

Claims

1. A method for determining a local intent of a query comprising:

receiving a query;

identifying a presence of a location term in the query;

separating the query into a root term and the location term when the location term is present;

analyzing historical searches that record which location terms are most often associated with the root term; and

computing a local intent probability (LIP) for the root term with a probability model using the historical searches wherein computing the LIP comprises a probability model using the historical searches, wherein the probability model is an exponential function between the LIP and a local intent value (LIV).

2. The method of claim 1 wherein historical searches are broken into root terms that are associated with location bands, wherein the location bands cover different geographic ranges.

3. The method of claim 2 wherein the LIP is calculated based on which location bands the query root term is associated with.

4. The method of claim 2 wherein the location bands comprise five location bands each covering a different geographic range.

5. The method of claim 4 wherein the geographic ranges of the five location bands includes a country, state, large city, small city, zip code, neighborhood, or combinations thereof.

6. The method of claim 1 wherein the probability model utilizes a truth set from the historical searches.

7. The method of claim 1 further comprising:

associating location bands with the root term based on the location terms in the historical searches that are associated with the root term;

assigning a value for each of the location bands, wherein the respective values represent a geographic range;

multiplying a frequency the root term is associated with one of the location bands with the value of the one of the location bands; and

combining the multiplied values for the root term to form the LIV.

8. In a computer readable storage medium having stored therein data representing instructions executable by a programmed processor for providing a local intent probability, the storage medium comprising instructions operative to:

receiving a plurality of search queries;

separating a root term and a location term from each of the plurality of queries;

calculating the local intent probability for the root terms from the plurality of queries based on the frequency each root term is associated with the location terms;

categorizing the location terms into location bands each covering a different geographic range;

computing the frequency in each of the location bands for the root terms from the plurality of queries; and

determining the probability that the root terms appear in each of the location bands.

9. The storage medium of claim 8 wherein each root term from the plurality of search queries is associated with at least one of the location terms.

10. The storage medium of claim 8 wherein the root term includes a frequency for each of the location bands, wherein the frequency in each of the location bands represents a number of times in the plurality of queries the root term was associated with a location term in each of the location bands.

11. The storage medium of claim 8 wherein the location bands comprise at least five location bands.

12. The storage medium of claim 11 wherein the geographic ranges of the five location bands includes a country, state, large city, small city, zip code, neighborhood, or combinations thereof.

13. The storage medium of claim 8 wherein the local intent probability represents a probability that a search query is geographically dependent.

14. The storage medium of claim 8 wherein the local intent probability is calculated from a distribution histogram of the location bands for a particular query.

15. A system determining a local intent probability (LIP) of a search query comprising:

a search engine coupled with a network and configured to receive the search query over the network;

a local intent analyzer coupled with the search engine and configured to extract a root term from the search query; and

a root term database coupled with the local intent analyzer and configured to store a plurality of root terms extracted from a plurality of historical search queries combined with at least one location associated with the plurality of historical root terms;

wherein the local intent analyzer determines the local intent of the query root term based on the at least one location associated with the query root term by computing the LIP with a regression model using the historical search queries stored in the root term database.

16. The system according to claim 15 further comprising a matching engine coupled with the local intent analyzer and configured to compare the query root term and the local intent of the query root term with available content to select relevant content.

17. The system according to claim 16 further comprising a user device coupled with the network and configured to receive the relevant content over the network in response to submitting the search query.

18. The system according to claim 15 wherein the local intent is expressed as a local intent probability that reflects a probability the search query is associated with a location.

19. The system according to claim 15 further comprising at least one of an internal data source or an external data source coupled with the root term database and configured to provide the plurality of search queries.

20. The system according to claim 15 further comprising a root extractor coupled with the local intent analyzer and configured to extract the query root term from the search query for the local intent analyzer.

21. The system according to claim 15 wherein the regression model comprises an exponential function between the LIP and a local intent value (LIV).