Systems and methods for inferring concepts for association with content

• US 7,673,253 B1
• Filed: 06/30/2004
• Issued: 03/02/2010
• Est. Priority Date: 06/30/2004
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Carobus, Alex, Ghosemajumder, Shuman, Axe, Brian, Ranganath, Rama, Agarwal, Sumit, Rajaram, Gokul
• Primary Examiner(s): Lo; Weilun
• Assistant Examiner(s): Chuong; Truc T
• Application Number:
  US10/879,521
• Time in Patent Office: 2,071 Days
• Field of Search: 715/760, 715/764, 715/789, 715/711, 715/715, 715/745, 715/811, 707/3, 707/5, 707/10
• US Class Current: 715/811
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/9558 Details of hyperlinks; Management of linked annotations

Abstract

A system indirectly infers concepts associated with a document. The concepts may be indirectly inferred based on information that does not include characteristics of the document, such as the characteristics that include a textual content of the document not associated with links included in the document, a domain of the document, and the document'"'"'s Uniform Resource Locator (URL). The system may label the inferred concepts as useful to an audience of the document.

Claims

1. A method, performed by one or more server devices, comprising:

retrieving, by a processor of the one or more server devices, a document;

indirectly inferring, by a processor of the one or more server devices, concepts associated with the document, where the concepts are not directly inferred from characteristics of the document and where indirectly inferring concepts associated with the document comprises;

identifying, by a processor of the one or more server devices, a plurality of second documents that each include a link pointing to the document,

determining, by a processor of the one or more server devices, which second document of the plurality of the second documents has been frequently used to access the document,

extracting, by a processor of the one or more server devices, concepts associated with the second document that has been frequently used to access the document,

determining, by a processor of the one or more devices, which advertisements contained in the document have been accessed from the document more than a threshold number of times,

extracting, by a processor of the one or more devices, a uniform resource locator (URL) associated with each of the advertisements that have been accessed from the document more than the threshold number of times,

retrieving, by a processor of the one or more devices, third documents corresponding to each of the URLs, and

extracting, by a processor of the one or more devices, concepts from each of the third documents; and

associating, by a processor of the one or more server devices, the inferred concepts with the document.

2. The method of claim 1, where the characteristics comprise a textual content of the document not associated with links included in the document, a domain of the document, and the document'"'"'s Uniform Resource Locator (URL).

3. The method of claim 1, where the document comprises at least one of a web page, an e-mail, a file, a combination of files, one or more files with embedded links to other files, a news group posting, a blog, or a web advertisement.

4. The method of claim 1, where indirectly inferring concepts associated with the document further comprises:

extracting a uniform resource locator (URL) associated with an advertisement contained in the document;

retrieving a third document corresponding to the URL; and

extracting concepts from the third document.

5. The method of claim 1, where indirectly inferring concepts associated with the document further comprises:

identifying at least one link in the document pointing from the document to another document; and

extracting concepts associated with the other document.

6. The method of claim 1, where indirectly inferring concepts associated with the document further comprises:

identifying links in the document pointing from the document to other documents;

determining which of the other documents have been accessed from the document more than a threshold number of times; and

extracting concepts associated with the other documents that have been accessed from the document more than the threshold number of times.

7. The method of claim 1, where indirectly inferring concepts associated with the document further comprises:

tracking user retrieval of a plurality of documents, where the plurality of documents include the document;

identifying third documents that the user retrieves before or after retrieval of the document; and

extracting concepts from each of the identified third documents.

8. The method of claim 1, where indirectly inferring concepts associated with the document further comprises:

extracting first concepts associated with at least one advertisement associated with the document;

determining advertisement campaigns that contain the extracted first concepts; and

extracting the concepts from the determined advertisement campaigns, where the concepts are different than the first concepts.

9. A computer-readable memory device containing instructions for controlling at least one processor to perform a method of inferring concepts associated with a document, the method comprising:

retrieving the document;

indirectly inferring concepts associated with the document, where indirectly inferring concepts associated with the document comprises;

identifying a plurality of second documents that each include a link pointing to the document,

determining which second document of the plurality of second documents has been most frequently used to access the document,

extracting concepts associated the second document that has been most frequently used to access the documents,

determining which advertisements contained in the document have been accessed from the document more than a threshold number of times,

extracting a uniform resource locator (URL) associated with each of the advertisements that have been accessed from the document more than the threshold number of times,

retrieving third documents corresponding to each of the URLs, and

extracting concepts from each of the third documents; and

associating the inferred concepts with the document.

10. The computer-readable memory device of claim 9, where the characteristics comprise a textual content of the document not associated with links included in the document, a domain of the document, and the document'"'"'s Uniform Resource Locator (URL).

11. The computer-readable memory device of claim 9, where the document comprises at least one of a web page, an e-mail, a file, a combination of files, one or more files with embedded links to other files, a news group posting, a blog, or a web advertisement.