Learning management server using Bayesian methods for adaptive hypermedia

1. A learning management system, comprising:

a processor;

a content storage unit in signal communication with the processor for storing learning content;

a user modeling unit in signal communication with the processor and having a user model, the user modeling unit comprising a macro level modeler and a micro level modeler, a user filter for distinguishing user types, and a user modeler in signal communication with the user filter for continuously inferring user sophistication level and user interest level in response to user interaction with the learning management system;

a personalization unit in signal communication with the processor for personalizing the learning content stored in the content storage unit in response to the user model; and

a user interface in signal communication with the processor for enabling a user to interact with the learning management system,

wherein the learning management system delivers content responsive to user interaction with the learning management system, and

wherein the user sophistication level is inferred using a data term indicative of the likelihood that a particular page sequence is viewed given a profile, and a prior term indicative of the likelihood of the profile itself.

2. A learning management system as defined in claim 1 configured as a Sharable Content Object Reference Model (“SCORM”) system.

3. A learning management system as defined in claim 1 wherein the user sophistication level is modeled using Bayesian Classification.

4. A learning management system as defined in claim 1 wherein the user modeler includes a Markov model to keep track of user navigation history.

5. A learning management system as defined in claim 4 wherein the Markov model is implemented using a finite order Markov chain.

6. A learning management system as defined in claim 1 wherein the user filter includes selection means for the selection of the user type.

7. A learning management system as defined in claim 6 wherein the selection means includes means for using a user description to select the appropriate user type.

8. A computer-implemented method of modeling a user to deliver the appropriate content, comprising:

modeling the user at a macro level by selecting a user type;

modeling the user at a micro level in response to the macro modeling by gauging the user sophistication and interest level; and

delivering the appropriate content to the user in accordance with the macro level modeling and the micro level modeling,

wherein the user sophistication level is inferred using a data term indicative of the likelihood that a particular page sequence is viewed given a profile, and a prior term indicative of the likelihood of the profile itself.

9. A method as defined in claim 8 , further comprising updating the model at the micro level in accordance with current user interactions.

10. A method as defined in claim 8 wherein the macro model is used to select the content to which the user will have access.

11. A method as defined in claim 8 wherein the user type is selected by analyzing the user's response to at least one question.

12. A method as defined in claim 8 , the step of modeling the user at the micro level comprising:

tracking user interaction;

tracking page visitation history; and

assigning a profile to the tracked user interaction and the tracked page visitation history.

13. A program storage device readable by machine, tangibly embodying a program of instructions executable by the machine to perform program steps for user modeling and content delivery, the program steps comprising:

modeling the user at a macro level by selecting a user type;

modeling the user at a micro level in response to the macro modeling by at least one of gauging the user sophistication and interest level; and

delivering the appropriate content to the user in accordance with the macro level modeling and the micro level modeling,

wherein the user sophistication level is inferred using a data term indicative of the likelihood that a particular page sequence is viewed given a profile, and a prior term indicative of the likelihood of the profile itself.

14. A device as defined in claim 13 , the program steps further comprising updating the model at the micro level in accordance with current user interactions.

15. A device as defined in claim 13 wherein the macro model is used to select the content to which the user will have access.

16. A device as defined in claim 13 wherein the user type is selected by analyzing the user's response to at least one question.

17. A device as defined in claim 13 , the program step of modeling the user at the micro level comprising:

tracking user interaction;

tracking page visitation history; and

assigning a profile to the tracked user interaction and the tracked page visitation history.

18. A device as defined in claim 13 , the program steps further comprising:

storing learning content;

personalizing the stored learning content in response to the user model; and

delivering content responsive to the user model.