MACHINE LEARNING FOR OPTIMAL STUDENT GUIDANCE

• US 20200134759A1
• Filed: 09/10/2019
• Published: 04/30/2020
• Est. Priority Date: 10/30/2018
• Status: Active Grant

• Current Assignee: Oracle International Corporation (Oracle Corporation)
• Original Assignee: Oracle International Corporation (Oracle Corporation)
• Inventors: Refila, John Edward, Thompson, Christopher Stanwood, Dreszer, Desiree Carvalho, Osibin, Nyeri Leah, Lauria, Michael Rene, Thompson, Alexander Winfield
• Granted Patent: US 11,423,501 B2
• Time in Patent Office: Days
• CPC Class Codes: G06F 17/18 for evaluating statistical data , e.g. average values, frequency distributions, probability functions, regression analysis forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04
• CPC Class Codes: G06F 18/217 Validation; Performance evaluation; Active pattern learning techniques
• CPC Class Codes: G06F 18/22 Matching criteria, e.g. proximity measures
• CPC Class Codes: G06F 18/23 Clustering techniques
• CPC Class Codes: G06F 18/24 Classification techniques
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06Q 50/2053 Education institution selection, admissions, or financial aid

Abstract

Techniques for training and evaluating machine-learning models for providing student guidance are described herein. In some embodiments, a network service generates a set of clusters that group a plurality of students by similarity. The network service trains a machine-learning model based on variances in outcomes and actions leading to the outcomes for students that belong to a same cluster in the set of clusters. The network service evaluates the machine-learning model for a student that has been mapped to the same cluster to identify at least one action that the student has not performed that is predictive of an optimal outcome for other students that belong to the same cluster. Responsive to evaluating the machine-learning model, the network service presents, through an interface, a recommendation that the student perform the at least one action.

Claims

1. One or more non-transitory computer-readable medium storing instructions which, when executed by one or more hardware processors, causes:

generating, by a network service, a set of clusters that group a plurality of students by similarity;

training, by the network service, a machine-learning model based on variances in outcomes and actions leading to the outcomes for students that belong to a same cluster in the set of clusters;

evaluating, by the network service, the machine-learning model for a student that has been mapped to the same cluster in the set of clusters to identify at least one action that the student has not performed that is predictive of an optimal outcome for other students that belong to the same cluster;

responsive to evaluating the machine-learning model, presenting, by the network service through an interface, a recommendation that the student perform the at least one action.

2. The one or more non-transitory computer-readable media of claim 1, wherein generating the set of clusters that group students by similarity comprises generating a feature vector for each student of the plurality of students;

assigning a respective student to a cluster in the plurality of clusters based on the distance of the feature vector to a cluster centroid.

3. The one or more non-transitory computer-readable media of claim 1, wherein the machine-learning model is one of a classification model or a regression model.

4. The one or more non-transitory computer-readable media of claim 1, wherein the at least one action comprises applying for a particular source of financial aid for which the student is predicted to be eligible.

5. The one or more non-transitory computer-readable media of claim 1, wherein the at least one action comprises adjusting an amount of financial aid accepted from a particular source of fund.

6. The one or more non-transitory computer-readable media of claim 5, wherein the at least one action includes reducing funds accepted from a first set of one or more sources and/or increasing funds accepted from a second set of one or more sources.

7. The one or more non-transitory computer-readable media of claim 1, wherein the at least one action comprises accessing a resource that is predicted to increase an academic outcome for the student.

8. The one or more non-transitory computer-readable media of claim 1, wherein the instructions further comprise presenting a link for performing the at least one action;

responsive to a user selecting the link navigating to an interactive interface through which the at least one action may be completed.

9. The one or more non-transitory computer-readable media of claim 7, wherein the link initiates an application process for applying for a loan.

10. The one or more non-transitory computer-readable media of claim 1, wherein the optimal outcome is determined, based at least in part, on input received from an adviser.

11. The one or more non-transitory computer-readable media of claim 1, wherein the optimal outcome is determined automatically, based at least in part, on a set of rules and/or criteria for identifying optimal outcomes.

12. The one or more non-transitory computer-readable media of claim 1, wherein the optimal outcome is determined, based at least in part, on a student debt burdens for students in the same cluster.

13. The one or more non-transitory computer-readable media of claim 1, wherein the at least one action is recommended based on a learned pattern of behavior that leads to reduced student debt for students in the same cluster.

14. The one or more non-transitory computer-readable media of claim 1, wherein the at least one action is presented through at least one of a student interface or an adviser interface.

15. The one or more non-transitory computer-readable media of claim 1, wherein the machine-learning model is evaluated in real-time as a student is accessing the interface.

16. The medium of claim 1, wherein the instructions further cause monitoring, by the network service, for triggers that cause changes to current or projected outcomes for the plurality of students;

and updating the machine-learning model based on detected changes to the current or projected outcomes for the plurality of students.

17. The one or more non-transitory computer-readable media of claim 16, wherein monitoring for triggers comprises receiving, by the network service from a network host, a RESTful message;

parsing the RESTful message to identify updated student data for at least one student in the plurality of students;

determining whether any outcomes are changed based on the updated student data;

responsive to determining that an outcome is changed based on the updated student data, classifying the updated student data as a trigger.

18. The medium of claim 17, wherein the RESTful message is a SAFI message that comprises a snapshot of current attributes of a particular student.

19. A system comprising:

one or more hardware processors;

one or more non-transitory computer-readable media storing instructions which, when executed by one or more hardware processors, causes;

generating, by a network service, a set of clusters that group a plurality of students by similarity;

training, by the network service, a machine-learning model based on variances in outcomes and actions leading to the outcomes for students that belong to a same cluster in the set of clusters;

evaluating, by the network service, the machine-learning model for a student that has been mapped to the same cluster in the set of clusters to identify at least one action that the student has not performed that is predictive of an optimal outcome for other students that belong to the same cluster;

responsive to evaluating the machine-learning model, presenting, by the network service through an interface, a recommendation that the student perform the at least one action.

20. A method comprising

generating, by a network service, a set of clusters that group a plurality of students by similarity;

training, by the network service, a machine-learning model based on variances in outcomes and actions leading to the outcomes for students that belong to a same cluster in the set of clusters;

evaluating, by the network service, the machine-learning model for a student that has been mapped to the same cluster in the set of clusters to identify at least one action that the student has not performed that is predictive of an optimal outcome for other students that belong to the same cluster;

responsive to evaluating the machine-learning model, presenting, by the network service through an interface, a recommendation that the student perform the at least one action.