Generating ranked search results using linear and nonlinear ranking models

• US 9,471,643 B2
• Filed: 06/04/2014
• Issued: 10/18/2016
• Est. Priority Date: 06/19/2009
• Status: Active Grant

• Current Assignee: Alibaba Group Holding Ltd.
• Original Assignee: Alibaba Group Holding Ltd.
• Inventors: Guo, Ning, Xing, Fei, Xie, Yuheng, Hou, Lei, Zhang, Qin
• Primary Examiner(s): Coby, Frantz
• Application Number:
  US14/296,191
• Publication Number:
  US 20140351246A1
• Time in Patent Office: 867 Days
• Field of Search: 707/723, 707/706, 707/752, 707/748
• US Class Current: 1/1
• CPC Class Codes: G06F 16/24578 using ranking
• CPC Class Codes: G06F 16/248 Presentation of query results
• CPC Class Codes: G06F 16/30 of unstructured textual data document management systems G06F16/93
• CPC Class Codes: G06F 16/438 Presentation of query results
• CPC Class Codes: G06F 16/44 Browsing; Visualisation therefor
• CPC Class Codes: G06F 16/93 Document management systems
• CPC Class Codes: G06F 16/957 Browsing optimisation, e.g. caching or content distillation

Abstract

Generating ranked search results includes receiving a plurality of matching information items that match a search request, ranking at least some of the plurality of matching information items using a linear ranking model that linearly combines a first plurality of feature values to obtain a first set of ranked results, ranking at least some of the first set of ranked results using a nonlinear ranking model that nonlinearly combines a second plurality of feature values to obtain a second set of ranked results, and provide a search response based on the second set of ranked results.

Claims

1. A system, comprising:

one or more processors configured to;

obtain a first set of ranked results by using a first ranking model to determine first rank scores corresponding to respective ones of at least some of a plurality of matching information items, wherein the first ranking model comprises a linear model;

select a portion of the first set of ranked results based at least in part on one or more selection criteria, wherein the one or more selection criteria comprises;

a number of search results to display at a page and a ratio of a selected number of the first set of ranked results to a total number of the first set of ranked results;

obtain a second set of ranked results by using a second ranking model to determine second rank scores corresponding to respective ones of at least some of the selected portion of the first set of ranked results, wherein the second ranking model comprises a nonlinear ranking model; and

generate a search response based at least in part on the second set of ranked results; and

one or more memories coupled to the one or more processors and configured to provide instructions to the one or more processors.

2. The system of claim 1, wherein the plurality of matching information items comprises a second plurality of matching information items and wherein the one or more processors are further configured to:

receive a search request;

generate a first plurality of matching information items based at least in part on the search request; and

select from the first plurality of matching information items the second plurality of matching information items.

3. The system of claim 2, wherein the first plurality of matching information items includes a greater number of matching information items than the second plurality of matching information items.

4. The system of claim 1, wherein determining the first rank scores corresponding to the respective ones of the at least some of the plurality of matching information items includes linearly combining a first plurality of feature values associated with the at least some of the plurality of matching information items.

5. The system of claim 4, wherein the first plurality of feature values is determined by one or both of the plurality of matching information items and a search request.

6. The system of claim 1, wherein to determine the second rank scores corresponding to the respective ones of the at least some of the selected portion of the first set of ranked results includes nonlinearly combining a second plurality of feature values associated with the at least some of the selected portion of the first set of ranked results.

7. The system of claim 1, wherein the second ranking model is determined by training.

8. The system of claim 1, wherein the first ranking model includes a constant or a first degree term.

9. The system of claim 1, wherein the second ranking model includes a term that is not associated with a constant or a first order term.

10. The system of claim 1, wherein the first set of ranked results includes a greater number of results than the second set of ranked results.

21. The system of claim 1, wherein:

the obtaining of the first set of ranked results by using the first ranking model that comprises the linear model causes the corresponding determined first rank scores to have a linear relationship with one or more feature values associated with the at least some of the plurality of matching information items; and

the obtaining of the second set of ranked results by using the second ranking model that comprises the nonlinear ranking model causes the corresponding determined second rank scores to have a nonlinear relationship with one or more feature values associated with the at least some of the selected portion of the first set of ranked results.

11. A method, comprising:

obtaining a first set of ranked results by using a first ranking model to determine first rank scores corresponding to respective ones of at least some of a plurality of matching information items, wherein the first ranking model comprises a linear model;

selecting a portion of the first set of ranked results based at least in part on one or more selection criteria, wherein the one or more selection criteria comprises;

a number of search results to display at a page and a ratio of a selected number of the first set of ranked results to a total number of the first set of ranked results;

obtaining a second set of ranked results by using a second ranking model to determine second rank scores corresponding to respective ones of at least some of the selected portion of the first set of ranked results, wherein the second ranking model comprises a nonlinear ranking model; and

generating a search response based at least in part on the second set of ranked results.

12. The method of claim 11, wherein the plurality of matching information items comprises a second plurality of matching information items and the method further comprising:

receiving a search request;

generating a first plurality of matching information items based at least in part on the search request; and

selecting from the first plurality of matching information items the second plurality of matching information items.

13. The method of claim 12, wherein the first plurality of matching information items includes a greater number of matching information items than the second plurality of matching information items.

14. The method of claim 11, wherein determining the first rank scores corresponding to the respective ones of the at least some of the plurality of matching information items includes linearly combining a first plurality of feature values associated with the at least some of the plurality of matching information items.

15. The method of claim 14, wherein the first plurality of feature values is determined by one or both of the plurality of matching information items and a search request.

16. The method of claim 11, wherein determining the second rank scores corresponding to the respective ones of the at least some of the selected portion of the first set of ranked results includes nonlinearly combining a second plurality of feature values associated with the at least some of the selected portion of the first set of ranked results.

17. The method of claim 11, wherein the second ranking model is determined by training.

18. The method of claim 11, wherein the first ranking model includes a constant or a first degree term.

19. The method of claim 11, wherein the second ranking model includes a term that is not associated with a constant or a first order term.

20. The method of claim 11, wherein the first set of ranked results includes a greater number of results than the second set of ranked results.