SYSTEM AND METHODS FOR EVALUATING INFERENCES OF UNKNOWN ATTRIBUTES IN A SOCIAL NETWORK

• US 20080294589A1
• Filed: 05/21/2008
• Published: 11/27/2008
• Est. Priority Date: 05/22/2007
• Status: Active Grant

• Current Assignee: Regents of the University of California (University of California)
• Original Assignee: Regents of the University of California (University of California)
• Inventors: Chu, Wesley W., Liu, Zhenyu, He, Jianming
• Granted Patent: US 8,160,993 B2
• Time in Patent Office: Days
• US Class Current: 706/55
• CPC Class Codes: G06F 21/6245 Protecting personal data, e.g. for financial or medical purposes

Abstract

A method and system for determining a probability of inferring an unknown attribute value for an attribute of interest for a target node in a social network. The method comprises the steps of receiving as an input attribute values and link relationships for a plurality of additional nodes in the social network, creating a simplified network using the input values and relationships, and calculating the probability of inferring the unknown attribute value for the target node.

Claims

1. A method of determining a probability of inferring an unknown attribute value for an attribute of interest for a target node in a social network, the method comprising the steps of:

receiving as an input attribute values and link relationships for a plurality of additional nodes in the social network;

creating a simplified network using the input values and relationships; and

calculating the probability of inferring the unknown on attribute value for the target node.

2. The method of claim 1, wherein the target node includes a plurality of direct descendents and indirect descendents, and the simplified network excludes any indirect descendent whose parent has a known value for the attribute of interest.

3. The method of claim 1, wherein the target node includes a plurality of direct descendents, and links between direct descendents are ignored during the step of creating the simplified network.

4. The method of claim 1, wherein the unknown attribute value is assumed, during the step of creating the simplified network, to influence the attribute values for the attribute of interest for all descendents of the target node.

5. The method of claim 1, wherein the social network is assumed, during the step of creating the simplified network, to be homogeneous.

6. The method of claim 1, wherein impacts of additional attributes on the attribute of interest are not considered.

7. A method of protecting the privacy of users in a social network, the method comprising the steps of:

determining a probability of inferring an unknown attribute value for an attribute of interest for a target user in the social network;

determining whether the probability is below or above an acceptable threshold; and

if the probability is above the acceptable threshold, alerting the target user.

8. The method of claim 7, further comprising taking steps to reduce the probability.

9. The method of claim 8, wherein the steps include modifying the social network.

10. A system for protecting the privacy of users in a social network, comprising:

a network creator module configured to receive as an input attribute values and link relationships for a plurality of nodes in the social network, and to create a simplified network using the input values and relationships; and

an inference engine configured to receive as an input from the network creator module the attribute values and link relationships, to calculate a plurality of conditional probability values, and to infer a conditional probability value for inferring an unknown attribute value for a target node.

11. The system of claim 10, further comprising an alert manager module configured to provide an alert to a member of the social network corresponding to the target node, if the inferred conditional probability value is greater than a threshold value.

12. The system of claim 10, further comprising a change monitor module configured to receive as an input from the network creator module a structure of the social network and the simplified network, to monitor changes to the social network over time, to compare the changes to the social network, and to communicate the changes to the network creator module.

13. The system of claim 11, further comprising:

a recommendation manager module; and

a network modifier module;

wherein the recommendation manager module is configured to receive signals from the alert manager module and to provide signals to the network modifier module.

14. The system of claim 13, wherein the network modifier module is further configured to simulate changes to the social network to create a modified social network, and to feed the simulated changes to the inference engine.

15. The system of claim 14, wherein the inference engine is further configured to receive as an input from the network modifier module attribute values and link relationships for the nodes in the modified social network, to calculate a plurality of conditional probability values, and to infer a conditional probability value for inferring the unknown attribute of interest for the target node.

16. The system of claim 15, wherein the recommendation manager module is configured to receive the inferred conditional probability value based upon the modified social network.

17. The system of claim 16, wherein the recommendation manager module is configured to initiate, based upon the inferred conditional probability value, one or more requests to the network modifier module to further simulate a change to the social network and to calculate an additional conditional probability value for inferring the unknown attribute of interest for the target node.

18. The system of claim 17, wherein the simulated change made to the social network is selected from the group consisting of at least one of (a) adding one or more additional nodes with one or more selected attribute values, (b) deleting one or more nodes with one or more selected attribute values, (c) changing an attribute value of an existing node, (d) adding one or more links to nodes with one or more selected attribute values, and (e) deleting one or more links to nodes with one or more selected attribute values.

19. The system of claim 17, wherein the recommendation manager module is further configured to compile changes made to the social network and corresponding conditional probability value for inferring the unknown attribute of interest for the target node.

20. The system of claim 19, wherein the recommendation manager module is further configured to recommend changes to be initiated by a user of the social network corresponding to the target node to maintain the conditional probability value for inferring the unknown attribute of interest for the target node below a threshold.