METHODS AND SYSTEMS FOR UTILIZING CONTENT, DYNAMIC PATTERNS, AND/OR RELATIONAL INFORMATION FOR DATA ANALYSIS

• US 20070118498A1
• Filed: 11/22/2006
• Published: 05/24/2007
• Est. Priority Date: 11/22/2005
• Status: Active Grant

• Current Assignee: NEC Corporation
• Original Assignee: NEC Laboratories America Inc (NEC Corporation)
• Inventors: Tseng, Belle, Song, Xiaodan
• Granted Patent: US 7,853,485 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/337 Profile generation, learning or modification
• CPC Class Codes: G06F 16/34 Browsing; Visualisation therefor
• CPC Class Codes: G06Q 30/0633 Lists, e.g. purchase orders, compilation or processing

Abstract

The present invention is directed generally to providing systems and methods for data analysis. More specifically, embodiments may provide system(s) and method(s) including dynamic user modeling techniques to capture the relational and dynamic patterns of information content and/or users'"'"' or entities'"'"' interests. Various embodiments may include system(s) and method(s) that are based on, for example, the past history of content semantics, temporal changes, and/or user community relationship. Various embodiments may include modeling and/or analysis of the dynamic nature of an item of interest'"'"'s value to a user(s)/entity(ies) over time. The dynamic factors may be consider in any manner, such as, individually or combined, sequentially or simultaneously, etc. Further, some embodiments may include, for example, system(s) and method(s) relating to analyzing data to capture user/entity interests and/or characteristics, consider content semantics and evolutionary information, and/or using community relationships of users/entities to thereby analyze information and provide dynamic conclusion(s) (e.g., recommendation(s)).

Claims

1. A method of dynamic pattern analysis, comprising the steps of:

detecting or determining one or more property(ies) of an item(s); and

associating a user(s) interest with the one or more property(ies) of an item.

2. The method of claim 1, wherein the one or more property(ies) includes an item'"'"'s life span.

3. The method of claim 1, wherein the one or more property(ies) includes expiration of data.

4. The method of claim 1, wherein the one or more property(ies) includes popularity of an item.

5. The method of claim 1, wherein the one or more property(ies) includes freshness of an item.

6. The method of claim 1, further comprising the step of predicting user(s) interests from the item properties.

7. The method of construct a community comprising:

determining a topic(s); and

determining a community(ies) based on the topic(s) and time.

8. The method of claim 7, wherein the community(ies) is an informal community or a formal community.

9. The method of claim 8, wherein the formal community is determined based on an organization chart and the informal community is determined by analyzing the topic(s) and time.

10. A method of discriminating data, comprising:

(this is the learning method) valuing more recent data differently than older data.

11. The method of claim 10, wherein the step of valuing includes applying a weight factor to indicate relative importance of the data.

12. A data analysis system, comprising:

a data set including user information, item information, and log files;

a content analysis module coupled to the data set; and

a community construction module coupled to the data set.

13. The system of claim 12, further comprising a learning module.

14. The system of claim 13, wherein the learning module includes time-sensitivity.

15. The system of claim 14, wherein the learning module includes an Adaboost procedure.

16. The system of claim 13, further comprising a recommender module coupled to the learning module, wherein the recommender module outputs a prediction of recommendation.

17. The system of claim 12, further comprising a dynamic pattern analysis module.

18. The system of claim 17, wherein the dynamic pattern analysis module includes analysis for determining popularity of an item, expiration status of an item, life span of an item, freshness of an item, and/or user intention.

19. A method of data analysis, comprising the steps of:

determining dynamic patterns over time of user characteristics;

determining dynamic patterns over time of information of interest to users; and

generating predictions or recommendations based on the users'"'"' dynamic pattern variations and the documents'"'"' dynamic patterns.

20. The method of claim 19, wherein the dynamic patterns over time of user characteristics includes user intention and change in user interest.

21. The method of claim 19, wherein dynamic patterns over time of information of interest to users includes long-term information and short-term information grouping and information expiration date prediction.

22. The method of claim 21, wherein the information item is a document.

23. The method of claim 19, further comprising the step of creating formal communities of users and/or informal communities of users using time as a factor.

24. The method of claim 23, wherein a content-time-community model creates the informal communities.

25. The method of claim 19, further comprising the step of providing an adaptive user model to generate the predictions or recommendations.

26. The method of claim 25, wherein the adaptive user model is a stage-wise estimation procedure for fitting an additive logistic regression model by minimizing a loss function.

27. The method of claim 25, wherein the adaptive user model is a time-sensitive Adaboost procedure.