Method and system for ranking web pages in a search engine based on direct evidence of interest to end users

• US 8,442,974 B2
• Filed: 06/24/2009
• Issued: 05/14/2013
• Est. Priority Date: 06/27/2008
• Status: Active Grant

• Current Assignee: Walmart Apollo, LLC (WalMart Inc.)
• Original Assignee: Wal-Mart Stores Texas LLC (WalMart Inc.)
• Inventors: Young, David L., Pavlidis, Ioannis, Watler, William Randall, Musk, Kimbal J., Newcomb, Peter J., Reich, Robert
• Primary Examiner(s): HASAN, SYED HAROON
• Assistant Examiner(s): HASAN, SYED HAROON
• Application Number:
  US12/491,104
• Publication Number:
  US 20090327281A1
• Time in Patent Office: 1,420 Days
• Field of Search: None
• US Class Current: 707/723
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation

Abstract

A method and system for ranking Web pages in a Web search engine is described. One illustrative embodiment receives a Web search query from a particular user, the query including at least one keyword; identifies one or more Web pages that contain the at least one keyword; determines, for each of the one or more Web pages, a raw page ranking; adjusts the raw page ranking of each of at least one Web page among the one or more Web pages based on direct evidence of how interesting that Web page is to users to produce an adjusted page ranking, the direct evidence being derived from clickstream data collected from the users; and presents, as search results, the at least one Web page to the particular user in accordance with the adjusted page rankings.

Claims

1. A computer-server-based method for ranking Web pages in a Web search engine, the method comprising:

receiving, over a network at the computer server, a Web search query from a particular user, the Web search query including at least one keyword, the computer server hosting the Web search engine;

identifying one or more Web pages that contain the at least one keyword;

determining, for each of the one or more Web pages, a raw page ranking;

adjusting the raw page ranking of each of at least one Web page among the one or more Web pages based on direct evidence of how interesting the respective Web page is to users to produce an adjusted page ranking, the direct evidence being derived from clickstream data collected from the users; and

presenting, as search results, the at least one Web page to the particular user in accordance with the adjusted page rankings,

wherein the direct evidence of how interesting a Web page is to users includes a measure of how often the users traverse to or from the Web page in browsing the Web and a measure of how many users have recently visited the Web page compared with how many users normally visit the Web page, and

wherein the measure of how many users have recently visited the Web page compared with how many users normally visit the Web page is given greater weight than the measure of how often the users traverse to or from the Web page in browsing the Web.

2. The computer-server-based method of claim 1, wherein adjusting the raw page ranking of each of at least one Web page among the one or more Web pages based on direct evidence of how interesting the respective Web page is to users to produce an adjusted page ranking includes scaling a number of user traversals to the respective Web page from other Web pages, as indicated by the clickstream data.

3. The computer-server-based method of claim 1, wherein the measure of how often the users traverse to or from a Web page in browsing the Web includes a correlation coefficient that is strengthened by additional user traversals to the Web page from other Web pages and weakened by an absence, over time, of additional user traversals to the Web page from other Web pages.

4. The computer-server-based method of claim 1, wherein the measure of how many users have recently visited a Web page compared with how many users normally visit the Web page is based on how many users visited the Web page during a recent predetermined unit of time compared with how many users, on average, visited the Web page during one or more predetermined units of time preceding the recent predetermined unit of time, as indicated by the clickstream data.

5. The computer-server-based method of claim 1, wherein the raw page ranking of each of the at least one Web page is adjusted in real time in response to the Web search query.

6. The computer-server-based method of claim 1, wherein the raw page ranking of each of the at least one Web page is based, at least in part, on how frequently the at least one keyword occurs in the respective Web page.

7. The method of claim 1, wherein the measure of how many users have recently visited the Web page compared with how many users normally visit the Web page is given greater weight than the measure of how often the users traverse to or from the Web page in browsing the Web by a factor of at least 10.

8. A system for ranking Web pages in a Web search engine, the system comprising:

at least one processor;

a communication interface; and

a memory containing a plurality of program instructions configured to cause the at least one processor to;

receive, via the communication interface over a network, a Web search query from a particular user, the Web search query including at least one keyword;

identify one or more Web pages that contain the at least one keyword;

determine, for each of the one or more Web pages, a raw page ranking;

adjust the raw page ranking of each of at least one Web page among the one or more Web pages based on direct evidence of how interesting the respective Web page is to users to produce an adjusted page ranking, the direct evidence being derived from clickstream data collected from the users; and

present, as search results, the at least one Web page to the particular user in accordance with the adjusted page rankings,

wherein the direct evidence of how interesting a Web page is to users includes a measure of how often the users traverse to or from the Web page in browsing the Web and a measure of how many users have recently visited the Web page compared with how many users normally visit the Web page, and

wherein the measure of how many users have recently visited the Web page compared with how many users normally visit the Web page is given greater weight than the measure of how often the users traverse to or from the Web page in browsing the Web.

9. The system of claim 8, wherein the plurality of program instructions are configured to cause the at least one processor to adjust the raw page ranking of each of at least one Web page among the one or more Web pages based on direct evidence of how interesting the respective Web page is to users to produce an adjusted page ranking by performing calculations that include scaling a number of user traversals to the respective Web page from other Web pages, as indicated by the clickstream data.

10. The system of claim 8, wherein the measure of how often the users traverse to or from a Web page in browsing the Web includes a correlation coefficient that is strengthened by additional user traversals to the Web page from other Web pages and weakened by an absence, over time, of additional user traversals to the Web page from other Web pages.

11. The system of claim 8, wherein the measure of how many users have recently visited a Web page compared with how many users normally visit the Web page is based on how many users visited the Web page during a recent predetermined unit of time compared with how many users, on average, visited the Web page during one or more predetermined units of time preceding the recent predetermined unit of time, as indicated by the clickstream data.

12. The system of claim 8, wherein the plurality of program instructions are configured to cause the at least one processor to adjust the raw page ranking of each of the at least one Web page in real time in response to the Web search query.

13. The system of claim 8, wherein the raw page ranking of each of the at least one Web page is based, at least in part, on how frequently the at least one keyword occurs in the respective Web page.

14. The system of claim 8, wherein the measure of how many users have recently visited the Web page compared with how many users normally visit the Web page is given greater weight than the measure of how often the users traverse to or from the Web page in browsing the Web by a factor of at least 10.

15. A non-transitory computer-readable medium having stored thereon computer-executable instructions that when executed by a computer cause the computer to:

receive, over a network at the computer server, a Web search query from a particular user, the Web search query including at least one keyword, the computer server hosting a Web search engine;

identify one or more Web pages that contain the at least one keyword;

determine, for each of the one or more Web pages, a raw page ranking;

access direct evidence of how interesting the respective Web page is to users, the direct evidence including data derived from clickstream data collected from the users, a measure of how often the users traverse to or from the respective Web page in browsing the Web, and a measure of how many users have recently visited the respective Web page compared with how many users normally visit the respective Web page;

adjust the raw page ranking of each of at least one Web page among the one or more Web pages based on the direct evidence to produce an adjusted page ranking, wherein the measure of how many users have recently visited the respective Web page compared with how many users normally visit the respective Web page is given greater weight than the measure of how often the users traverse to or from the respective Web page in browsing the Web; and

present, as search results, the at least one Web page to the particular user in accordance with the adjusted page rankings.

16. The non-transitory computer-readable medium of claim 15, wherein the computer-executable instruction for adjusting the raw page ranking includes scaling a number of user traversals to the respective Web page from other Web pages, as indicated by the clickstream data.

17. The non-transitory computer-readable medium of claim 15, wherein the measure of how often the users traverse to or from a Web page in browsing the Web includes a correlation coefficient that is strengthened by additional user traversals to the Web page from other Web pages and weakened by an absence, over time, of additional user traversals to the Web page from other Web pages.

18. The non-transitory computer-readable medium of claim 15, wherein the measure of how many users have recently visited a Web page compared with how many users normally visit the Web page is based on how many users visited the Web page during a recent predetermined unit of time compared with how many users, on average, visited the Web page during one or more predetermined units of time preceding the recent predetermined unit of time, as indicated by the clickstream data.

19. The non-transitory computer-readable medium of claim 15, wherein the raw page ranking of each of the at least one Web page is adjusted in real time in response to the Web search query.

20. The non-transitory computer-readable medium of claim 15, wherein the raw page ranking of each of the at least one Web page is based, at least in part, on how frequently the at least one keyword occurs in the respective Web page.