Systems and methods for determining user attribute values by mining user network data and information

• US 8,965,974 B2
• Filed: 08/17/2012
• Issued: 02/24/2015
• Est. Priority Date: 08/19/2011
• Status: Active Grant

• Current Assignee: Board of Regents of the University of Texas System (University of Texas System)
• Original Assignee: Board of Regents of the University of Texas System (University of Texas System)
• Inventors: Abrol, Satyen, Khadilkar, Vaibhav, Khan, Latifur Rahman, Thuraisingham, Bhavani Marienne
• Primary Examiner(s): Golabbakhsh, Ebrahim
• Application Number:
  US13/588,977
• Publication Number:
  US 20130218965A1
• Time in Patent Office: 921 Days
• Field of Search: 709/204
• US Class Current: 709/204
• CPC Class Codes: H04L 67/306 User profiles
• CPC Class Codes: H04L 67/52 specially adapted for the location of the user terminal
• CPC Class Codes: H04L 67/535 Tracking the activity of the user network monitoring arrangements H04L43/00; recording of computer activity G06F11/34

Abstract

Systems and methods for determining one or more attributes and their associated values for a user, by mining the user'"'"'s social network profiles. The systems and methods determine if the value of an attribute of interest for a user is specified on one or more social networking websites. If so, the systems and methods set the attribute value for the user to the specified attribute value. If not, the systems and methods retrieve, from the social networking websites, attribute values for the attribute of interest for friends of the user to form a group of possible attribute values for the user. The method, selects a value from the group of possible attribute values, and then sets the selected value as the attribute value for the user.

Claims

1. A method for determining a value of an attribute of a user, comprising:

a processor in the data processing system determining if the value of the attribute of interest for the user is specified on one or more social networking websites;

responsive to a determination that the value of an attribute of the user is not specified, the processor obtaining, from the one or more social networking websites, attribute values for the attribute of interest for friends of the user to form a group of possible attribute values for the user, each of the possible attribute values associated with a concept;

the processor assigning a weight to each of the possible attribute values based on a specificity of each of the possible attribute values, wherein the specificity is based on a number of words in each of the possible attribute values;

the processor selecting a value from the group of possible attribute values based on the weight; and

the processor setting the selected value as the value of the attribute for interest for the user.

2. The method of claim 1, further comprising:

responsive to a determination that the value of the attribute of the user is specified, the processor retrieving and setting a specified value as the value of the attribute of interest for the user.

3. The method of claim 1, wherein the attribute of interest for the user comprises one or more of a location, age, age group, race, ethnicity, threat, languages spoken, religion, economic status, education level, gender, hobby, or interest.

4. The method of claim 1, wherein the processor determining if a value of an attribute of interest for the user is specified on one or more social networking websites further comprises:

the processor obtaining an identifier of the user from a social network profile of the user; and

the processor obtaining the value of the attribute of interest for the user, if specified, using the identifier of the user.

5. The method of claim 1, wherein the processor obtaining the attribute values corresponding to the attribute of interest for the friends of the user to form the group of possible attribute values for the user comprises:

the processor obtaining a user identifier for each of the friends of the user from their social network profiles; and

the processor obtaining the attribute values for the friends, if specified on the one or more social networking websites, using the user identifier of each friend.

6. The method of claim 1, wherein the processor selecting the value from the group of possible attribute values comprises:

the processor selecting the value having a maximum frequency among the friends of the user.

7. The method of claim 1, wherein the processor selecting the value from the group of possible attribute values comprises:

the processor determining a set of closest friends of the user;

the processor obtaining a value for the attribute of interest for each friend in the set of closest friends;

the processor aggregating and determining a weight for each obtained value; and

the processor selecting the value, from the obtained values for the set of closest friends, having a highest weighted score.

8. The method of claim 7, wherein the processor determining a set of closest friends of the user further comprises:

the processor determining a number of friends in common between the user and each friend of the user; and

the processor determining the set of closest friends based on the number of friends in common between the user and each friend.

9. The method of claim 8, further comprising:

the processor determining a friend of the user is a spammer if a difference between a number of users the spammer follows and a number of users following the spammer is equal to or greater than a threshold; and

responsive to determining the friend of the user is a spammer, the processor setting the number of friends in common between the user and the friend to zero (0).

10. The method of claim 1, wherein the processor selecting the value from the group of possible attribute values further comprises:

the processor determining a set of closest friends of the user;

the processor obtaining a value for the attribute of interest for each friend in the set of closest friends, wherein the obtained values are stored as an array and wherein each obtained value in the array is associated with a probability score; and

the processor selecting the value from the array of values having a highest calculated score among the set of closest friends.

11. The method of claim 10, wherein the attribute is a location, and further comprising:

combining two concepts, wherein each concept represents a location;

the processor assigning one or more weights to each concept;

the processor calculating a weight score for each concept based on the weights of each concept; and

wherein the combining, assigning, and calculating steps are repeated until a determination is made that the combined concepts have a weight score greater than a location confidence threshold (LCT).

12. The method according to claim 11, wherein the combined concepts represent a potential location vector.

13. The method of claim 1, wherein the value of the attribute of interest for the user is specified on a publicly available profile of the one or more social networking websites.

14. A method in a data processing system for determining a location of a user, comprising:

extracting text from messages provided by a user on a social networking website, wherein the messages include one or more location concepts;

removing words unrelated to geographical locations from the text;

assigning a weight to each location concept based on the occurrence of the location concept in the text, wherein specific location concepts are assigned a greater weight compared to general location concepts;

responsive to determining a correlation between a specific location concept and a general location concept, adding the weight of the general location concept to the specific location concept;

reordering the location concepts in descending order of weights;

assigning a probability score (S) to each location concept (C) based on its individual weight; and

based on the probability scores, assigning a location concept to the user.

15. The method of claim 14, wherein the location concept comprises a specified geographical location and is of the format {City}/{State}/{Country}.

16. The method of claim 15, wherein a value of one or more of the City, the State, or the Country is null.

17. An apparatus for determining a value of an attribute of a user, comprising:

a bus;

a storage device connected to the bus, wherein the storage device contains computer usable code; and

a processing unit connected to the bus, wherein the processing unit executes the computer usable code to;

determine if a value of an attribute of interest for a user is specified on one or more social, networking websites;

obtain, from the social networking websites, attribute values for the attribute of interest for direct connections of the user to form a group of possible attribute values for the user in response to a determination that the value of the attribute of the user is not specified, wherein the direct connections of the user are friends of the user;

obtain, from the social networking websites, attribute values for the attribute of interest for indirect connections of the user to form the group of possible attribute values for the user in response to a determination that the attribute value of one of the direct connections of the user is not specified, wherein the indirect connections of the user are only linked to the one of the direct connections of the user;

select a value from the group of possible attribute values; and

set the selected value as the value of the attribute of interest for the user.

18. The apparatus of claim 17, wherein the selection of the value from the group of-possible attribute values further comprises a selection of the value from the group of possible attribute values having a maximum frequency among the direct connections and indirect connections of the user.

19. The apparatus of claim 17, wherein the selection of value from the group of possible attribute values further comprises:

a determination of a number of friends in common between the user and each friend of the user;

a determination of a set of closest friends based on the number of friends in common between the user and each friend;

obtaining a value for the attribute of interest for each friend in the set of closest friends;

aggregating and determining a weight for each obtained value; and

a selection of a value from the obtained values having a highest weighted score among the set of-closest friends.

20. The apparatus of claim 17, the selection of value from the group of possible attribute values further comprises:

a determination of a number of friends in common between the user and each friend of the user;

a determination of a set of closest friends based on the number of friends in common between the user and each friend;

obtaining a value for the attribute of interest for each friend in the set of closest friends, wherein the obtained, values comprise an array of values, and wherein each value in the array is associated with a probability score; and

a selection of a value from the array of values having a highest calculated score among the set of closest friends.