Techniques for mobile prediction

• US 9,852,375 B2
• Filed: 12/26/2014
• Issued: 12/26/2017
• Est. Priority Date: 12/26/2014
• Status: Active Grant

• Current Assignee: Intel Corporation
• Original Assignee: Intel Corporation
• Inventors: Sanjay, Addicam V., Avalos, Jose A., Jensen, Joe D.
• Primary Examiner(s): Holmes, Michael B
• Application Number:
  US14/583,524
• Publication Number:
  US 20160189038A1
• Time in Patent Office: 1,096 Days
• Field of Search: 706 55
• CPC Class Codes: G06N 5/04 Inference or reasoning models

Abstract

An apparatus and method for mobile prediction is disclosed. In an embodiment, a predictive component creates graphs of user behaviors and receives subgraphs derived from a global graph. By combining a subset of the user behavior graph and the subgraph, a user behavior is predicted. Other embodiments are described and claimed.

Claims

1. A mobile device comprising:

a control circuit;

a storage unit communicatively coupled to the control circuit, the storage unit to store a prediction component operative on the control circuit to;

create a user activity graph based on a set of user information and a set of mobile device information;

receive a global subgraph from a server coupled to the mobile device via a network; and

derive a prediction of user activity based on combining at least a portion of the user activity graph and the global subgraph.

2. The mobile device of claim 1, the global subgraph to comprise one of a set of global subgraphs created from a global graph.

3. The mobile device of claim 2, the global graph to comprise one or more nodes derived from product-keyword associations.

4. The mobile device of claim 3, the global subgraph to comprise at least one edge and multiple nodes, where a distance between the nodes is normalized.

5. The mobile device of claim 4, the normalized distance to be used to select the global subgraph.

6. The mobile device of claim 1, comprising a memory, transceiver, one or more antennas, and a display.

7. The mobile device of claim 1, the storage unit to store a user interface component operative on the control circuit to present the prediction of user activity on an electronic display.

21. The mobile device of claim 1, the global subgraph comprising indications of a plurality of user activity graphs, each of the plurality of user activity graph based on different sets of user information and different sets of mobile device information.

22. The mobile device of claim 21, the set of user information comprises indications of at least one of the user'"'"'s name, a user'"'"'s browsing history, a quantity of emails the user receives on the mobile device, a quantity of emails the user sends on the mobile device, a quantity of text messages the user receives on the mobile device, a quantity of text messages the user sends of the mobile device, or a shopping history of the user on the mobile device.

23. The mobile device of claim 21, the set of mobile device information comprises indications of at least one of a temperature, an ambient light level, a location, system setting, or an installed application.

8. A method for predicting a user behavior, comprising:

creating a user activity graph based on a set of user information and a set of mobile device information;

receiving, via a network, a global subgraph from a server, the global subgraph to be stored in a database accessible to the server; and

deriving a prediction of user activity based on combining at least a portion of the user activity graph and the global subgraph.

9. The method of claim 8, the global subgraph to comprise one of a set of global subgraphs created from a global graph.

10. The method of claim 9, the global graph to comprise one or more nodes derived from product-keyword associations.

11. The method of claim 10, the global subgraph to comprise one edge and multiple nodes, where a distance between the nodes is normalized.

12. The method of claim 8, the global subgraphs being based on the normalized distance.

13. The method of claim 8, comprising displaying the user prediction on a display.

14. A non-transient computer readable storage medium containing executable instructions to:

create a user activity graph based on a set of user information and a set of mobile device information;

receive, via a network, a global subgraph from a server, the global subgraph to be stored in a database accessible to the server; and

derive a prediction of user activity based on combining at least a portion of the user activity graph and the global subgraph.

15. The non-transient computer readable storage medium of claim 14, the global graph to comprise one or more nodes derived from product-keyword associations.

16. The non-transient computer readable storage medium of claim 15, the global subgraph to comprise at least one edge and multiple nodes, where a distance between the nodes is normalized.

17. The non-transient computer readable storage medium of claim 16, the normalized distance to be used to select the global subgraph.

18. The non-transient computer readable storage medium of claim 14, further comprising instructions to present the user activity on an electronic display.

19. The non-transient computer readable storage medium of claim 18, the prediction of user activity to include a product recommendation.

20. The non-transient computer readable storage medium of claim 18, the prediction of user activity to include a non-product recommendation.