FAST ADAPTIVE DOCUMENT FILTERING

• US 20090198683A1
• Filed: 02/27/2009
• Published: 08/06/2009
• Est. Priority Date: 05/15/2003
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Robertson, Stephen, Walker, Stephen, Zobel, Justin
• Granted Patent: US 8,095,559 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/335 Filtering based on additional data, e.g. user or group profiles filtering in web context G06F16/9535, G06F16/9536
• CPC Class Codes: G06F 16/337 Profile generation, learning or modification
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access
• CPC Class Codes: Y10S 707/99936 Pattern matching access
• CPC Class Codes: Y10S 707/99943 Generating database or data structure, e.g. via user interface

Abstract

Data structures, stored on various types of computer-readable media, include information related to user profiles and/or to various documents. The information included in these data structures is arranged and stored in manner that allows for rapid user profile updating to be performed as new or changed documents are processed in a document filtering system.

Claims

9. (canceled)

19. One or more computer-readable storage medium having computer-executable instructions which, when executed by one or more processors, configures the processors to perform a method, the method comprising:

creating a document statistics file that indicates a frequency of various terms in a new document, without regard to any of a plurality of user profiles;

updating an original user profile of the plurality of user profiles based at least in part on terms from the new document and feedback from a user corresponding to the original user profile, the original user profile specifying one or more areas of interest to the user, the feedback regarding relevance of a prior document, the feedback existing prior to determining whether the new document is relevant to any of the plurality of user profiles;

updating a reference dictionary file data structure, the reference dictionary file data structure including terms parsed from a plurality of documents stored in a document repository, at least one term parsed from the new document not stored in the document repository; and

updating a parsed term data structure, the parsed term data structure including a one or more parsed terms and a term selection value associated with each of the one or more parsed terms, each of the parsed terms either;

present in the reference dictionary file data structure, indicating at least one document indicated as relevant to the user contains the term, or

present in at least one of the plurality of user profiles,

at least one of the parsed terms present in at least one of the plurality of user profiles and not present in the reference dictionary file data, the term selection value used for determining whether the associated term is to be included in the updated user profile.

1. A computer-readable storage medium having computer-executable instructions for performing acts, the acts comprising:

creating a document statistics file that indicates a frequency of various terms in a document;

determining a profile score using the document statistics file and information associated with a user profile;

notifying a user associated with the user profile if the profile score exceeds a profile threshold value;

updating a reference dictionary file to include terms from the document; and

updating a document dictionary index to include terms from the document.

2. A computer-readable storage medium as defined in claim 1, wherein reference dictionary file is stored on a first computer-readable media and the document dictionary index is stored on a second computer-readable media.

3. A computer-readable storage medium as defined in claim 1, wherein reference dictionary file is stored in Random Access Memory (RAM) and the document dictionary index is stored on a mass storage device.

4. A computer-readable storage medium as defined in claim 1, wherein the reference dictionary file includes terms located in a plurality of documents stored in a document repository and terms located in the document, but not in the plurality of documents.

5. A computer-readable storage medium as defined in claim 1, wherein the reference dictionary file includes terms located in a plurality of documents stored in a document repository and terms not located in the plurality of documents and the document dictionary index includes terms located only in the plurality of documents.

6. A computer-readable storage medium as defined in claim 5, the acts further comprising creating on the first computer-readable media a parsed term data structure including each of the terms in the reference dictionary file, a plurality of term weights, and a plurality of term selection values, wherein each term weight and term selection value is associated with a separate one of the terms in the parsed term data structure.

7. A computer-readable storage medium as defined in claim 5, the acts further comprising creating a parsed term data structure including each of the terms in the reference dictionary file, and updating a user profile based on the parsed term data structure.

8. A computer-readable storage medium as defined in claim 1, the acts further comprising:

retrieving one or more documents;

retrieving a user profile associated with the user; and

forming a parsed term data structure including information derived from the received one or more documents, the retrieved user profile and the reference dictionary file.

10. A computer-readable storage medium as defined in claim 7, the parsed term data structure further comprising an original profile indicator associated with each of the parsed terms in the parsed term data structure, each original profile indicator indicating whether a parsed term associated therewith is present in a specified user profile.

11. A computer-readable storage medium as defined in claim 7, the parsed term data structure further comprising a term weight associated with each term in the parsed term data structure, and a term selection value associated with each term in the parsed term data structure, the term selection value a product of the term weight and a total number of relevant documents in which the term appears.

12. A computer-readable storage medium as defined in claim 1, the acts further comprising:

retrieving one or more documents specified as relevant to a user;

retrieving a user profile associated with the user;

forming a parsed term data structure including information derived from the received one or more documents, the retrieved user profile and the reference dictionary file; and

updating the user profile based on the information in the parsed term data structure.

13. A computer-readable storage medium as defined in claim 1, wherein the information associated with a user profile is based at least in part on feedback from the user.

14. A computer-readable storage medium as defined in claim 13, wherein the feedback from the user is received directly from the user.

15. A computer-readable storage medium as defined in claim 1, wherein feedback from the user is inferred from the user'"'"'s action associated with the document.

16. A computer-readable storage medium as defined in claim 1, wherein feedback the information associated with a user profile is based at least in part on feedback received directly from the user.

17. A computer-readable storage medium as defined in claim 1, the information associated with the user profile updated based at least in part on terms from the document included in the reference dictionary file.

18. A computer-readable storage medium as defined in claim 1, the method further comprising:

reading the profile threshold value from a profile threshold record contained within a profile threshold data structure, the profile threshold data structure including a plurality of profile threshold records, each profile threshold record including a profile identifier field and a profile threshold field.

21. The one or more computer-readable storage medium of claim 19, wherein the feedback from the user is based at least in part from the user'"'"'s action associated with the new document.

21-1. The one or more computer-readable storage medium of claim 19, wherein the feedback from the user is based at least in part from feedback received directly from the user.