Web query classification

• US 9,424,346 B2
• Filed: 04/23/2012
• Issued: 08/23/2016
• Est. Priority Date: 01/28/2005
• Status: Active Grant

• Current Assignee: Mercury Kingdom Assets Limited
• Original Assignee: Mercury Kingdom Assets Limited
• Inventors: Chowdhury, Abdur R., Beitzel, Steven Michael, Lewis, David Dolan, Kolcz, Aleksander
• Primary Examiner(s): Rostami, Mohammad S
• Assistant Examiner(s): ABRAHAM, AHMED M
• Application Number:
  US13/453,901
• Publication Number:
  US 20120209870A1
• Time in Patent Office: 1,583 Days
• Field of Search: 707/766
• US Class Current: 1/1
• CPC Class Codes: G06F 16/3331 Query processing
• CPC Class Codes: G06F 16/353 into predefined classes
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques

Abstract

A query phrase may be automatically classified to one or more topics of interest (e.g., categories) to assist in routing the query phrase to one or more appropriate backend databases. A selectional preference query classification technique may be used to classify the query phrase based on a comparison between the query phrase and patterns of query phrases. Additionally, or alternatively, a combination of query classification techniques may be used to classify the query phrase. Topical classification of a query phrase also may be used to assist a search system in delivering auxiliary information to a user who entered the query phrase. Advertisements, for instance, may be tailored based on classification rather than query keywords.

Claims

1. A computer-implemented method, comprising:

receiving a query phrase;

parsing the query phrase into at least a first constituent part and a second constituent part;

determining, with at least one processor, a first category associated with the received query phrase by performing a first classification process that uses the first constituent part, definitional information associated with the first constituent part, and the second constituent part;

determining, with the at least one processor, a second category associated with the received query phrase by performing a second classification process that uses the first constituent part and the second constituent part, the second classification process being different than the first classification process;

accessing, with the at least one processor, a pattern that is associated with at least one of the first category or the second category;

determining whether the first category matches the second category;

in response to a determination that the first category does not match the second category, selecting at least one category from among at least the first category and the second category based at least in part on the pattern, wherein the selecting is based at least in part on an estimated strength of the first classification process in determining that the received query phrase is associated with the first category and an estimated strength of the second classification process in determining that the received query phrase is associated with the second category;

associating the received query phrase with the selected category; and

processing the query phrase based on the selected category.

2. The computer-implemented method of claim 1, the first classification process comprising:

comparing the first constituent part with a first part included in the accessed pattern; and

determining whether the first constituent part corresponds to the accessed pattern, based on the comparison results.

3. The computer-implemented method of claim 2, the first classification process further comprising:

comparing the second constituent part with a second part included in the accessed pattern;

determining whether the second constituent part corresponds to the accessed pattern, based on the comparison results; and

in response to determining that the first constituent part and the second constituent part correspond to the accessed pattern, identifying the at least one category that is associated with the pattern as the first category associated with the received query phrase.

4. The computer-implemented method of claim 1, further comprising:

accessing a first classification metric that indicates a confidence level in classifications made by the first classification process; and

accessing a second classification metric that indicates a confidence level in classifications made by the second classification process,

wherein the selecting at least one category further comprises selecting a category based on the first classification metric and the second classification metric.

5. The computer-implemented method of claim 1, further comprising:

determining, with the at least one processor, a third category associated with the received query phrase by performing a third classification process that uses the first constituent part and the second constituent part, the third classification process being different than the first classification process and the second classification process, wherein the determination of whether the first category matches the second category comprises determining whether at least one of the first category and second category matches the third category, and

wherein the selecting at least one category further comprises selecting a category from among the first category, second category, and third category, in response to a determination that not all of the first category, second category, and third category match.

6. The computer-implemented method of claim 5, the selecting at least one category further comprising:

identifying a category determined by a majority of the classification processes and using the identified category as the selected category.

7. The computer-implemented method of claim 1, further comprising:

identifying, based on the selected category, at least one search resource for satisfying the received query phrase; and

processing the received query phrase by using the at least one search resource to generate at least one search result.

8. The computer-implemented method of claim 1, further comprising:

modifying the received query phrase by at least one of adding words to the received query phrase, eliminating words from the received query phrase, or reordering words in the received query phrase; and

processing the modified query phrase to generate at least one search result.

9. The computer-implemented method of claim 1, wherein the selecting at least one category further comprises selecting both the first category and the second category, wherein the association of the received query phrase with the selected category comprises associating the received query phrase with both the first category and the second category, and wherein the processing the query phrase based on the selected category comprises processing the query phrase based on both the first category and the second category.

20. The computer-implemented method of claim 1, further comprising determining the definitional information associated with the first constituent part, wherein the definitional information includes a description of the first constituent part.

21. The computer-implemented method of claim 1, wherein the processing the query phrase based on the selected category includes routing the query phrase to one or more backend databases based on the selected category.

22. The computer-implemented method of claim 21, wherein the one or more backend databases include at least one of a private database, a website, a bulletin board, or a backend search engine.

23. The computer-implemented method of claim 1, further comprising:

accessing a data store that includes a list of backend databases, the data store associating each of the backend databases with one or more categories; and

determining, from the list of backend databases, at least one backend database that is associated with the selected category;

wherein the processing the query phrase based on the selected category includes routing the query phrase to the at least one backend database that is associated with the selected category.

24. The computer-implemented method of claim 1, further comprising:

determining, from a plurality of backend databases that are individually associated with one or more categories, a set of backend databases that are associated with the selected category, the set of backend databases including at least a first backend database and a second backend database;

determining that the first backend database includes first information that is redundant with second information included in the second backend database; and

determining, from the set of backend databases that are associated with the selected category, a subset of backend databases that excludes the first backend database based on the determining that the first backend database includes the first information that is redundant with the second information included in the second backend database; and

wherein the processing the query phrase based on the selected category includes routing the query phrase to the subset of backend databases.

25. The computer-implemented method of claim 1, further comprising:

accessing a data store that includes a list of backend databases, the data store associating each of the backend databases with one or more categories; and

determining, from the list of backend databases, a particular backend database that is associated with the selected category;

wherein the processing the query phrase based on the selected category includes;

reformatting, based at least in part on a format corresponding to the particular backend database, the query phrase to produce a reformatted query phrase;

routing the reformatted query phrase to the particular backend database that is associated with the selected category.

10. An apparatus comprising:

at least one processor;

at least one storage medium having instructions stored thereon which, when executed by the at least one processor, cause the at least one processor to;

parse a received query phrase into at least a first constituent part and a second constituent part;

determine a first category associated with the received query phrase by performing a first classification process that uses the first constituent part, definitional information associated with the first constituent part, and the second constituent part;

determine a second category associated with the received query phrase by performing a second classification process that uses the first constituent part and the second constituent part, the second classification process being different than the first classification process;

access a pattern that is associated with at least one of the first category or the second category;

determine whether the first category matches the second category;

in response to a determination that the first category does not match the second category, select at least one category from among at least the first category and the second category based at least in part on the pattern, wherein the selecting is based at least in part on an estimated strength of the first classification process in determining that the received query phrase is associated with the first category and an estimated strength of the second classification process in determining that the received query phrase is associated with the second category;

associate the received query phrase with the selected category; and

process the query phrase based on the selected category.

11. The apparatus of claim 10, the at least one storage medium further including instructions that, when executed by the at least one processor, cause the at least one processor to:

compare the first constituent part with a first part included in the accessed pattern; and

determine whether the first constituent part corresponds to the accessed pattern, based on the comparison results.

12. The apparatus of claim 11, the at least one storage medium further including instructions that, when executed by the at least one processor, cause the at least one processor to:

compare the second constituent part with a second part included in the accessed pattern;

determine whether the second constituent part corresponds to the accessed pattern, based on the comparison results; and

in response to determining that the first constituent part and the second constituent part correspond to the accessed pattern, identify the at least one category that is associated with the pattern as the first category associated with the received query phrase.

13. The apparatus of claim 10, the at least one storage medium further including instructions that, when executed by the at least one processor, cause the at least one processor to:

access a first indication that indicates the estimated strength of the first classification process in determining that the received query phrase is associated with the first category; and

access a second indication that indicates the estimated strength of the second classification process in determining that the received query phrase is associated with the second category.

14. The apparatus of claim 10, the at least one storage medium further including instructions that, when executed by the at least one processor, cause the at least one processor to:

access a first classification metric that indicates a confidence level in classifications made by the first classification process; and

access a second classification metric that indicates a confidence level in classifications made by the second classification process,

wherein the selecting at least one category further comprises selecting a category based on the first classification metric and the second classification metric.

15. The apparatus of claim 10, the at least one storage medium further including instructions that, when executed by the at least one processor, cause the at least one processor to:

determine a third category associated with the received query phrase by performing a third classification process that uses the first constituent part and the second constituent part, the third classification process being different than the first classification process and the second classification process,

wherein the determination of whether the first category matches the second category comprises determining whether one of the first category and second category matches the third category, and

wherein the selecting at least one category further comprises selecting a category from among the first category, second category, and third category, in response to a determination that not all of the first category, second category, and third category match.

16. The apparatus of claim 15,

wherein the selecting at least one category further comprises identifying a category determined by a majority of the classification processes and using the identified category as the selected category.

17. The apparatus of claim 10, the at least one storage medium further including instructions that, when executed by the at least one processor, cause the at least one processor to:

identify, based on the selected category, at least one search resource for satisfying the received query phrase; and

process the received query phrase by using the at least one search resource to generate at least one search result.

18. The apparatus of claim 10, the at least one storage medium further including instructions that, when executed by the at least one processor, cause the at least one processor to:

modify the received query phrase by at least one of adding words to the received query phrase, eliminating words from the received query phrase, or re-ordering words in the received query phrase; and

process the modified query phrase to generate at least one search result.

19. The apparatus of claim 10, wherein the selecting at least one category further comprises selecting both the first category and the second category, wherein the association of the received query phrase with the selected category comprises associating the received query phrase with both the first category and the second category, and wherein the processing the query phrase based on the selected category comprises processing the query phrase based on both the first category and the second category.