SYSTEM AND METHOD FOR RECOMMENDING EDUCATIONAL RESOURCES

• US 20100159437A1
• Filed: 12/19/2008
• Published: 06/24/2010
• Est. Priority Date: 12/19/2008
• Status: Abandoned Application

• Current Assignee: PARC-Palo Alto Research Center, Palo Alto Research Center, Inc. (Xerox Holdings Corp.), Xerox Corporation (Xerox Holdings Corp.)
• Original Assignee: PARC-Palo Alto Research Center, Xerox Corporation (Xerox Holdings Corp.)
• Inventors: Perronnin, Florent, Zhou, Rong, Cifarelli, Claudio, Lofthus, Robert M., Do, Minh, German, Kristine A., Price, Robert Roy
• Application Number:
  US12/340,054
• Publication Number:
  US 20100159437A1
• Time in Patent Office: Days
• US Class Current: 434/433
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G09B 7/02 of the type wherein the student is expected to construct an answer to the question which is presented or wherein the machine gives an answer to the question presented by a student

Abstract

A recommender system and method is provided, including receiving a request to recommend a course of action related to a plurality of current students in accordance with a plurality of constraints and accessing a computer database storing student data that corresponds to the plurality of current students. The student data includes attribute data corresponding to respective students of the plurality of current students for describing at least one attribute related to the respective students. The method further includes clustering in a computer process the plurality of current students into a selected number of clusters based at least on sameness of attribute data corresponding to the respective current students of the plurality of current students and the plurality of constraints, and outputting the results of the clustering to a user.

Claims

1. A recommender system for recommending clustering of students into educational groups, the recommender system comprising:

a processor for executing a series of programmable instructions for;

receiving a request to cluster a plurality of current students into a selected number of educational groups in accordance with a plurality of constraints, wherein the selected number is at least two;

accessing student data corresponding to a plurality of students, wherein the student data includes attribute data corresponding to respective students of the plurality of students describing at least one attribute related to the student; and

clustering the plurality of current students into a selected number of clusters based at least on sameness of the attribute data corresponding to the respective current students of the plurality of current students and the plurality of constraints; and

outputting the results of the clustering to a user.

2. The recommender system according to claim 1, wherein the student data includes granular assessment data, wherein the granular assessment data includes a result of an assessment administered to a current student of the plurality of current students, wherein the assessment includes a plurality of questions for the assessing the current student and the result includes an independent evaluation of each respective question of the plurality of questions, wherein the clustering is further based on sameness of the granular assessment data associated with respective current students.

3. The recommender system in accordance with claim 1, wherein the granular assessment data is obtained from markings extracted from assessment answer sheets scanned in via a multifunctional device (MFD), wherein the assessment answer sheets are used for recording answers to questions associated with an assessment of the student.

4. The recommender system in accordance with claim 1, wherein the at least one student attribute corresponding to each respective student defines a location within a multidimensional space whose dimensions are each related to a respective attribute, wherein selected attributes of the at least one student attribute are each associated with a respective dimension of the multidimensional space, wherein the clustering based on sameness of the attribute data includes clustering based on a relationship between the locations associated with the respective current students.

5. The recommender system in accordance with claim 1, wherein the plurality of constraints include logistical constraints which must be satisfied by the clustering.

6. The recommender system in accordance with claim 5, wherein the clustering algorithm is a constraints based algorithm for satisfying a constraint satisfaction problem.

7. The recommender system in accordance with claim 1, wherein the respective constraints of the plurality of constraints have associated weights, and the clustering is further based on the weights of the associated constraints.

8. The recommender system in accordance with claim 7, the processor further executing programmable instructions for receiving adjustments from the user to at least one of the plurality of constraints and the associated weights;

and repeating the clustering, outputting and receiving adjustments until the user is satisfied with the output results.

9. The recommender system in accordance with claim 1, wherein the course of action in the request is to make classroom assignments for the plurality of current students.

10. The recommender system in accordance with claim 1, wherein the course of action in the request is to form the plurality of current students into individual groups for performing an educational activity.

11. The recommender system in accordance with claim 1, wherein the data storage facility further stores resource data including attributes related to a plurality of educational resources, the processor further executes programmable instructions for accessing the resource data;

and the course of action in the request includes recommending an educational resource for at least one of the clusters based on its related attributes.

12. The recommender system in accordance with claim 11, wherein the educational resource includes at least on of an educator, physical educational material, digital educational material and an instructional method.

14. The recommender system in accordance with claim 1, wherein the clustering is further based on a selected level of performance as indicated by the granular assessment data associated with at least one predecessor student of the plurality of students, which are students of the plurality of students, other than the current students, that have prior experience with a course of action related to the course of action in the request.

15. The recommender system in accordance with claim 14, wherein the student data further includes, associated with each student of the plurality of students, student attribute data describing at least one attribute of the student, and wherein the at least one predecessor student is selected based on sameness of the student data associated with the at least one predecessor student relative to the student data associated with the plurality of current students.

16. The recommender system in accordance with claim 15, wherein the student data associated with each respective student defines a location within a multidimensional space whose dimensions are each related with a respective attribute, wherein selected attributes of the at least one attribute re associated with a respective dimension of the multidimensional space, wherein the selection of the at least one predecessor student based on sameness of the student data includes selecting based on a relationship between the locations associated with respective students of the at least one predecessor student and respective current students of the plurality of current students.

17. The recommender system in accordance with claim 1, wherein the clustering includes using non-negative matrix factorization.

18. The recommender system in accordance with claim 1, wherein the clustering includes using probabilistic latent semantic analysis.

19. The recommender system in accordance with claim 1,

wherein each student of the plurality of students is associated with a D-dimensional vector encoding granular assessment data for that student, wherein the vector includes values indicative of an evaluation of each answer to D respective questions, wherein the value for each question of the D questions is +1 for a correct answer and −

1 for an incorrect answer; and

wherein the clustering includes determining the sameness of a first student having an associated vector u and a second student having an associated vector v, and applying a formula to u and v.

20. The recommender system in accordance with claim 19, wherein the formula is:

0.5*(1+cos(u,v)).

21. A method for recommending clustering of students into educational groups, the method comprising:

receiving a request to recommend a course of action related to a plurality of current students in accordance with a plurality of constraints;

accessing a computer database storing student data that corresponds to the plurality of current students, the student data including attribute data corresponding to respective students of the plurality of current students for describing at least one attribute related to the respective students;

clustering in a computer process the plurality of current students into a selected number of clusters based at least on sameness of attribute data corresponding to the respective current students of the plurality of current students and the plurality of constraints; and

outputting the results of the clustering to a user.

13. The recommender system in accordance with claim 32, wherein the educational resource includes natural language processing (NLP) reading material which is digitally modifiable to provide a reading level compatible with a reading level associated with the cluster it is recommended for, wherein the reading level associated with the cluster is determined based on the associated granular assessment data.

22. The method according to claim 21, wherein the clustering is performed in accordance with at least one user adjustable clustering constraint, wherein the at least one adjustable clustering constraint is selected from the group of clustering constraints consisting of:

a student attribute constraint specifying attribute data included in the student data describing attributes related to respective students of the plurality of students to use for determining sameness when the clustering is based on sameness, a differentiation constraint for specifying a degree of differentiation between clusters, a cluster size constraint specifying a minimum number of students to include in each cluster, a maximum number of students to include in each cluster, and the selected number of clusters, and an algorithm constraint specifying a clustering algorithm to apply.

23. The method according to claim 21, further comprising:

accessing resource data in the computer database describing attributes related to a plurality of educational resources;

recommending an educational resource for respective clusters of the clusters based on the resource data and the student data corresponding to students included in the respective cluster.

24. The method according to claim 23, wherein the student data includes data indicating needs or preferences of the respective students and the resource data includes data indicating strengths or capabilities of the respective resources, and the educational resource recommended for a particular clusters is recommended when resource data associated with the resource indicates that the resource'"'"'s strengths or capabilities are compatible with needs or preferences of the particular cluster indicated by the student data associated with the students included in the particular cluster.

25. The method according to claim 21, wherein the clustering is further based on a selected level of performance as indicated by the granular assessment data associated with at least one predecessor student of the plurality of students, which are students of the plurality of students, other than the current students, that have prior experience with a course of action related to the course of action in the request.

26. The method according to claim 25, wherein the at least one predecessor student is selected based on sameness of the student attribute data associated with the at least one predecessor student relative to the student attribute data associated with the plurality of current students.

27. The method in accordance with claim 26, wherein the student attribute data associated with each respective current and predecessor student defines a location within a multidimensional space whose dimensions are each related with a respective attribute, wherein selected attributes of the at least one attribute are each associated with a respective dimension of the multidimensional space, wherein the selection of the at least one predecessor student based on sameness of the student attribute data includes selecting based on a relationship between the locations associated with the respective current and predecessor students.

28. The method according to claim 27, wherein the at least one student attribute corresponding to each respective student defines a location within a multidimensional space whose dimensions are each related with a respective attribute, wherein selected attributes of the at least one student attribute are each associated with a respective dimension of the multidimensional space, wherein the clustering based on sameness of the attribute data includes clustering based on a relationship between the locations associated with the respective current students.

29. The method in accordance with claim 21, wherein the clustering includes using non-negative matrix factorization.

30. The method in accordance with claim 21, wherein the clustering includes using probabilistic latent semantic analysis.

31. The method in accordance with claim 21,

wherein each student of the plurality of students is associated with a D-dimensional vector encoding granular assessment data for that student, wherein the vector includes values indicative of an evaluation of each answer to D respective questions, wherein the value for each question of the D questions is +1 for a correct answer and −

1 for an incorrect answer; and

wherein the clustering includes determining the sameness of a first student having an associated vector u and a second student having an associated vector v, and applying a formula to u and v.

32. The method in accordance with claim 31, wherein the formula is:

0.5*(1+cos(u,v)).

33. A computer-readable medium storing a series of programmable instructions configured for execution by at least one processor for recommending clustering of students into educational groups comprising the steps of:

receiving a request to recommend a course of action related to a plurality of current students in accordance with a plurality of constraints;

accessing student data that corresponds to the plurality of current students, the student data including attribute data corresponding to respective students of the plurality of current students for describing at least one attribute related to the respective students;

clustering the plurality of current students into a selected number of clusters based at least on sameness of attribute data corresponding to the respective current students of the plurality of current students and the plurality of constraints; and

outputting the results of the clustering to a user.

34. The computer-readable medium according to claim 33, wherein respective constraints of the plurality of constraints have associated weights, and the clustering is further based on the weights of the associated constraints.

35. The computer-readable medium according to claim 33, wherein the programmable instructions further comprise the steps of:

receiving adjustments from the user to at least one of the plurality of constraints and the associated weights; and

continuing to repeat the clustering, outputting, and receiving adjustments until the user is satisfied with the output results.