Statistical models and methods to support the personalization of applications and services via consideration of preference encodings of a community of users

• US 20050132014A1
• Filed: 06/30/2004
• Published: 06/16/2005
• Est. Priority Date: 12/11/2003
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Horvitz, Eric J., Subramani, Murugesan S., Apacible, Johnson T.
• Granted Patent: US 7,774,349 B2
• Time in Patent Office: Days
• US Class Current: 709/206
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising

Abstract

A statistical methodology is employed for assisting people with the setting of parameters of software applications or services. With the method, information about demographics and/or about the complete or partial preferences of other people, are analyzed to infer recommendations for settings and functionality of a computer application or service. A system is reviewed with the ability to receive data regarding a plurality of users, and stores respective profiles for these users in a community store. The system can leverage off of these completed and/or partially completed profiles of parameters representing preferences about software operation in connection with building new profiles for users (new and/or existing). Data regarding a user that desires to build a new profile is employed in connection with the community profiles to facilitate the user building a personalized profile. Various statistical and/probabilistic schemes can be employed, for example, collaborative filtering techniques to identify to the user the top n settings by particular parameter(s), top m settings by popularity, top x most similar profiles to facilitate the user selecting most appropriate sub-profiles as part of a personalized profile building and selection effort.

Claims

1. A system that facilitates generation of a system profile, comprising:

a storage component that receives data relating to a community of users, and

an analyzer that processes the community data and information relating to existing profiles to provide recommendations to the user in connection with building a personalized profile.

2. The system of claim 1, the analyzer employs collaborative filtering to determine one or more profiles for a user.

3. The system of claim 1, the existing profiles are generated from a plurality of client computers that have generate profile settings in accordance with various applications.

4. The system of claim 3, the applications include communications systems, general computing system, operating system settings, development system applications, and applications employing configuration or operation settings that are applied to customize the applications according to users preferences.

5. The system of claim 1, the analyzer applies statistical processes, mathematical processes, or models to the storage component to determine one or more profiles that can be displayed and selected at a user interface.

6. The system of claim 5, the user interface includes various user inputs or selections to enable a user to guide selection of a desired profile via the analyzer.

7. The system of claim 1, a filter component to apply collaborative filtering in accordance with the analyzer to process previous system settings preferences of users to predict likely or possible settings or profiles for new users of a system.

8. The system of claim 7, the filter component includes one or more algorithms including memory based methods, model-based methods, and hybrid methods employing portions of the memory and model based methods.

9. The system of claim 7, the filter component employs explicit or implicit voting to determine potential profiles for a given user.

10. The system of claim 9, the filtering component determines votes of an active user based on partial information regarding the active user and a set of weights calculated from a user database.

11. The system of claim 7, the filter component employs at least one of a Bayesian classifier or a Bayesian network to determine a potential profile.

12. The system of claim 7, the filter component employs inference techniques or statistical clustering methods, to reason about a likelihood that a user with some preferences or demographic information would agree with other previously unspecified settings or preferences.

13. The system of claim 7, the filter component is employed for of identifying software settings or for identifying preferences about context-sensitive computing.

14. The system of claim 13, the filter component is employed for a definition of a cost of interruption in different settings relating to an application in communications.

15. The system of claim 14, the application processes context, including time of day, day of week, a call priority of individuals, and groups of people.

16. The system of claim 14, the application employs cost-benefit rules for routing messages between parties.

17. The system of claim 1, the analyzer determines a profile based upon demographic information of a potential user of the profile, according a popularity of existing profiles, according to similarity between profiles, and according to settings associated with a parameter.

18. The system of claim 1, further comprising a profile assistant interface to enable user to determine a desired profile.

19. The system of claim 18, the profile assistant is a web-based application.

20. A computer readable medium having computer readable instructions stored thereon for implementing the components of claim 1.

21. A method for automatically generating a preference profile, comprising:

storing a plurality of settings for a computer application in a database; and

automatically applying collaborative filtering to the database to determine a subset of preference settings for a respective user of the computer application, the preference settings employed to configure or operate the computer application.

22. A system for automatically generating a preference profile, comprising:

means for aggregating existing configurations for a computer or a computer application;

means filtering to the existing configurations to determine a potential subset of configuration settings for a respective user of the computer or the computer application; and

means for selecting or influencing the potential subset of configuration settings.

23. A graphical user interface to facilitate determine of preferences for an application, comprising:

a display object associated with viewing potential profiles relating to a community of users preferences;

at least one input to select or alter the potential profiles; and

a filter component to automatically determine the potential profiles.

24. The graphical user interface of claim 23, further comprising a My Profile option populated with the current settings that has been input to client software.

25. The graphical user interface of claim 23, further comprising a recommendations display of potential profiles determined via a collaborative filter.

26. The graphical user interface of claim 23, further comprising an input for selecting settings by similarity or by popularity.

27. The graphical user interface of claim 23, further comprising drag and drop capabilities for moving potential settings into a user profile setting.

28. The graphical user interface of claim 26, further comprising employing demographic information to determine the similarity.

29. The graphical user interface of claim 23, further comprising a starting point input to select a recommended profile.

30. The graphical user interface of claim 29, further comprising a component to update collaborative filtering profiles upon selection or a new profile or parameter.

31. The graphical user interface of claim 23, further comprising a profile display based upon a role or status of a user.