System and method for dynamically tracking user interests based on personal information

• US 8,838,588 B2
• Filed: 03/30/2005
• Issued: 09/16/2014
• Est. Priority Date: 03/30/2005
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Ruvolo, Joann, Edlund, Stefan Bengt, Krishna, Vikas, Lessler, Justin Thomas, Kraenzel, Carl J.
• Primary Examiner(s): Bullock, Joshua
• Application Number:
  US11/094,031
• Publication Number:
  US 20060230012A1
• Time in Patent Office: 3,457 Days
• Field of Search: 707/1, 707/732
• US Class Current: 707/732
• CPC Class Codes: G06F 16/337 Profile generation, learning or modification

Abstract

Described are a dynamic interest profile (DIP) system and method for dynamically tracking interests of a user based on personal information. The DIP system obtains electronic documents of the user from a document stream and processes the documents to obtain certain information therefrom. Based on the information obtained from the documents, the DIP system identifies terms, people, documents, and collections that are of importance to the user. These items of importance become part of a dynamic interest profile of the user. The dynamic interest profiles persist in a database. The DIP system also provides an application program interface (API) for accessing DIPs in the database. Application programs can employ this API to customize program behavior to the particular interests of the user executing those programs.

Claims

1. A method for tracking interests of users based on personal information, the method comprising:

monitoring a stream of documents from more than one user to obtain therefrom a plurality of electronic documents each belonging to one of the users;

generating, for each obtained electronic document, a dynamic interest profile (DIP) document based on information obtained from that electronic document;

associating each generated DIP document with one of the plurality of users;

uniquely assigning a document queue to each of the users;

selecting, for each DIP document, one of the document queues into which to place that DIP document based on the user with which that DIP document is associated;

polling each document queue to detect changes to that document queue;

determining, in response to detecting a change to the document queue of each user, at least one person of importance to that user and one term of importance to that user based on the DIP documents in the document queue uniquely assigned to that user;

generating an interest profile of each user that includes each person and each term of importance of that user; and

storing the interest profiles of the users in a DIP database.

2. The method of claim 1, further comprising determining from the information obtained from a given electronic document at least one item of emerging importance to the user to whom that electronic document belongs.

3. The method of claim 1, wherein the plurality of electronic documents includes a combination of email messages and word processing documents.

4. The method of claim 1, wherein generating, for each obtained electronic document, a dynamic interest profile (DIP) document from information obtained from that electronic document includes parsing each electronic document to identify terms and people mentioned in that electronic document.

5. The method of claim 1, further comprising determining a document of importance to a given user based on at least one term of importance to the given user and on at least one person of importance to the given user.

6. The method of claim 1, further comprising determining a collection of importance to a given user based on at least one term of importance to the given user and on at least one person of importance to the given user.

7. The method of claim 1, wherein determining at least one person of importance to each user and one term of importance to each user includes calculating an importance score for a term in the electronic documents belonging to that user based on at least one of a strength of association of the term with that user, a strength of association of the term with subjects of documents, a discriminative quality of the term for identifying contents of a folder, and a relevance of the term to that user.

8. The method of claim 1, wherein determining at least one person of importance to each user and one term of importance to each user includes calculating an importance score for a person mentioned in the electronic documents belonging to that user based on at least one of a longevity of association between the person and that user, when that user most recently interacted with the person, a number of bidirectional interactions with the person, a measure of exclusive interactions with the person, and a frequency of interactions with the person.

9. The method of claim 1, wherein determining at least one person of importance to each user and one term of importance to each user includes calculating an importance score for a document in the electronic documents belonging to that user based on at least one of a frequency of access by that user to the document, an importance of an author of the document, an importance of each person receiving the document, an importance of terms in a subject of the document, an importance of terms in a body of the document.

10. The method of claim 1, wherein determining at least one person of importance to each user and one term of importance to each user includes calculating an importance score for a collection of electronic documents belonging to that user based on at least one of a frequency of access by that user to the collection, a frequency of change by that user of electronic documents in the collection, and an importance of documents in the collection.

11. The method of claim 1, further comprising:

updating incremental feature vectors for each user based on the information obtained from the plurality of electronic documents for that user;

storing the incremental feature vectors for each user in the DIP database;

periodically performing an importance calculation using the incremental feature vectors for each user, to update the interest profile of that user; and

storing the importance feature vectors for each user in the DIP database.

12. The method of claim 1, wherein the stream of documents includes electronic documents generated by each user and electronic documents received by each user from someone other than that user.

13. The method of claim 1, further comprising:

providing an application program interface (API) for use by application programs to access the interest profiles of the users in the DIP database;

receiving a query issued by an application program using the API to access the interest profile of a given user in the DIP database; and

responding to the query by supplying to the application program an importance score vector having information computed to be of importance to the given user.

14. The method of claim 1, wherein the polling of each document queue occurs periodically at a rate controllable by a programmable processor.

15. The method of claim 1, further comprising determining from a DIP document in the document queue assigned to a given user if a new electronic document is being added to the electronic documents associated with the given user, if an existing electronic document associated with the given user is being modified, or if an existing electronic document associated with the given user is being deleted.

16. The method of claim 1, further comprising detecting deletion by a given user of an electronic document and removing, from the document queue assigned to the given user, the DIP document generated from that electronic document prior to deletion.

17. The method of claim 1, further comprising maintaining the DIP database on a server that is accessible to the users through client devices in communication with the server over a network.

18. The method of claim 1, further comprising tagging the electronic documents to indicate with which user each of the electronic documents is associated in order to facilitate placing each DIP document generated from each of the tagged electronic documents in an appropriate one of the document queues.

19. The method of claim 1, wherein a given one of the electronic documents is an email message having an email address representing a given user, and further comprising mapping the email address to a unique identifier of the given user in order to facilitate placing the DIP document generated from the given electronic document in the document queue assigned to the given user.

20. The method of claim 1, further comprising batch processing the DIP documents in the document queue uniquely assigned to a given user in response to polling that document queue.