Systems and methods to facilitate local searches via location disambiguation

• US 8,898,095 B2
• Filed: 06/24/2013
• Issued: 11/25/2014
• Est. Priority Date: 11/04/2010
• Status: Expired due to Fees

• Current Assignee: AT&T Intellectual Property I LP (AT&T, Inc.)
• Original Assignee: AT&T Intellectual Property I LP (AT&T, Inc.)
• Inventors: Agrawal, Ritesh Jitendra, Shanahan, James G.
• Primary Examiner(s): Gaffin, Jeffrey A
• Assistant Examiner(s): Misir, Dave
• Application Number:
  US13/924,750
• Publication Number:
  US 20130290227A1
• Time in Patent Office: 519 Days
• Field of Search: 706/12, 707/711
• US Class Current: 706/12
• CPC Class Codes: G06F 16/29 Geographical information databases
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06F 16/9537 Spatial or temporal dependent retrieval, e.g. spatiotemporal queries
• CPC Class Codes: G06N 20/00 Machine learning

Abstract

Systems and methods use machine learning techniques to resolve location ambiguity in search queries. In one aspect, a dataset generator generates a training dataset using query logs of a search engine. A training engine applies a machine learning technique to the training dataset to generate a location disambiguation model. A location disambiguation engine uses the location disambiguation model to resolve location ambiguity in subsequent search queries.

Claims

1. A method, comprising:

receiving, at a computing device from a searcher, a search query having location ambiguity;

generating, by the computing device, a location model using a machine learning technique, the location model generated, at least in part, based on a personalization feature, wherein the personalization feature comprises a location corresponding to a past result selected by the searcher, wherein generating the location model based, at least in part, on the personalization feature comprises generating a training dataset comprising the personalization feature, wherein generating the training dataset comprises

filtering a query log to identify a first query, the first query specifying an unambiguous location,

generating, from the first query, a second query having location ambiguity by removing a first component of the first query, wherein removing the first component comprises

breaking the first query into the first component and a second component, and

removing the first component,

identifying a location candidate for the second query, and

generating a training target based on comparing the unambiguous location specified in the first query and the location candidate for the second query;

identifying, by the computing device using the location model, an unambiguous location for the search query to resolve the location ambiguity; and

determining, by the computing device, a search result corresponding to the search query directed to the unambiguous location for the search query.

2. The method of claim 1, wherein the first component comprises a state and the second component comprises a city.

3. The method of claim 1, wherein the machine learning technique comprises a decision tree learning technique.

4. The method of claim 1, wherein the machine learning technique comprises a gradient boosted decision tree learning technique.

5. The method of claim 1, wherein the training dataset further comprises a location feature and a search feature.

6. The method of claim 5, wherein the location feature is independent of queries recorded in the query log.

7. The method of claim 5, wherein the location feature comprises at least one of:

area, population, population density, demography, popularity, or business population.

8. The method of claim 5, wherein the search feature is independent of searchers of the queries recorded in the query log.

9. The method of claim 5, wherein the search feature comprises at least one of:

topic-location co-occurrence frequency, time, day, or month.

10. The method of claim 1, wherein the personalization feature is related to searchers of queries recorded in the query log.

11. The method of claim 1, further comprising determining that the search query is received from the searcher via a mobile device,

wherein generating the location model comprises generating a first location model for queries received from mobile devices and generating a second location model for queries received from fixed-location devices, and

wherein identifying the unambiguous location for the search query using the location model comprises identifying the unambiguous location for the search query using the first location model based on the search query being received from the searcher via the mobile device.

12. The method of claim 1, wherein the personalization feature further comprises at least one of:

a location of the searcher, a direction being traveled by the searcher, or a speed being traveled by the searcher.

13. A non-transitory computer readable medium storing instructions which, when executed by a computer comprising a processor, cause the computer to perform operations comprising:

receiving a search query having location ambiguity;

generating a location model using a machine learning technique, the location model generated, at least in part, based on a personalization feature, wherein the personalization feature comprises a location corresponding to a past result selected by a searcher, wherein generating the location model based, at least in part, on the personalization feature comprises generating a training dataset comprising the personalization feature, wherein generating the training dataset comprises

filtering a query log to identify a first query, the first query specifying an unambiguous location,

generating, from the first query, a second query having location ambiguity by removing a first component of the first query, wherein removing the first component comprises

breaking the first query into the first component and a second component, and

removing the first component,

identifying a location candidate for the second query, and

generating a training target based on comparing the unambiguous location specified in the first query and the location candidate for the second query;

identifying, using the location model, an unambiguous location for the search query to resolve the location ambiguity; and

determining a search result corresponding to the search query directed to the unambiguous location for the search query.

14. The non-transitory computer readable medium of claim 13, wherein the operations further comprise determining that the search query is received from the searcher via a mobile device,

wherein generating the location model comprises generating a first location model for queries received from mobile devices and generating a second location model for queries received from fixed-location devices, and

wherein identifying the unambiguous location for the search query using the location model comprises identifying the unambiguous location for the search query using the first location model based on the search query being received from the searcher via the mobile device.

15. A system, comprising:

a processor; and

a memory that stores instructions that, when executed by the processor, cause the processor to perform operations comprising

receiving a search query having location ambiguity,

generating a location model using a machine learning technique, the location model generated, at least in part, based on a personalization feature, wherein the personalization feature comprises a location corresponding to a past result selected by a searcher, wherein generating the location model based, at least in part, on the personalization feature comprises generating a training dataset comprising the personalization feature, wherein generating the training dataset comprises

filtering a query log to identify a first query, the first query specifying an unambiguous location,

generating, from the first query, a second query having location ambiguity by removing a first component of the first query, wherein removing the first component comprises

breaking the first query into the first component and a second component, and

removing the first component,

identifying a location candidate for the second query, and

generating a training target based on comparing the unambiguous location specified in the first query and the location candidate for the second query,

identifying, using the location model, an unambiguous location for the search query to resolve the location ambiguity, and

determining a search result corresponding to the search query directed to the unambiguous location for the search query.

16. The system of claim 15, wherein the operations further comprise determining that the search query is received from the searcher via a mobile device,

wherein generating the location model comprises generating a first location model for queries received from mobile devices and generating a second location model for queries received from fixed-location devices, and

wherein identifying the unambiguous location for the search query using the location model comprises identifying the unambiguous location for the search query using the first location model based on the search query being received from the searcher via the mobile device.