Computing social influence scores for users

• US 8,977,617 B1
• Filed: 10/31/2012
• Issued: 03/10/2015
• Est. Priority Date: 10/31/2012
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Wu, Ping, Zhu, Kaihua
• Primary Examiner(s): Reyes, Mariela
• Assistant Examiner(s): Chang, Jeffrey
• Application Number:
  US13/664,621
• Time in Patent Office: 860 Days
• Field of Search: 707/736
• US Class Current: 707/736
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06Q 30/0201 Market modelling; Market analysis; Collecting market data
• CPC Class Codes: G06Q 30/0631 Item recommendations
• CPC Class Codes: G06Q 50/01 Social networking

Abstract

The disclosure includes a system and method for determining an influence score for a user. The system includes a classification module, a controller, a determination module, a ratio module and a score module. The classification module creates an influence profile for a first user. The controller determines activity data associated with a set of active users. The determination module determines how frequently the activity data associated with the active users describes the first user and a second user that has a second influence score. The determination module determines a first user activity score and a second user activity score. The ratio module compares the first user activity score to the second user activity score to determine an influence ratio. The score module determines a first influence score for the first user based at least in part on the influence ratio and the second influence score for the second user.

Claims

1. A method comprising:

determining whether an influence profile for a first user includes a first influence score on a social network;

responsive to an absence of the first influence score for the first user, determining activity data associated with a set of active users, the activity data including activity of the set of active users from one or more data sources independent of the social network;

determining a first frequency at which the activity data associated with the set of active users describes the first user and a second frequency at which the activity data associated with the set of active users describes a second user, the second user having an influence profile including a second influence score on the social network;

determining a first activity score for the first user based at least in part on the first frequency at which the activity data associated with the set of active users describes the first user, the first activity score for the first user indicating a first interest of the set of active users in the first user;

determining a second activity score for the second user based at least in part on the second frequency at which the activity data associated with the set of active users describes the second user, the second activity score for the second user indicating a second interest of the set of active users in the second user;

determining an influence ratio using the first activity score for the first user and the second activity score for the second user, the influence ratio corresponding to a ratio of the first influence score for the first user to the second influence score for the second user; and

determining the first influence score for the first user based at least in part on the influence ratio and the second influence score for the second user.

2. The method of claim 1, wherein determining the first frequency at which the activity data associated with the set of active users describes the first user further comprises:

determining a search query included in the activity data and a set of candidate users based at least in part on the search query;

matching the set of candidate users with one or more web page content elements associated with the search query to determine a set of matching scores for the set of candidate users;

determining a top candidate user from the set of candidate users based at least in part on the set of matching scores, the top candidate user having a highest matching score; and

determining whether the first user is the top candidate user.

3. The method of claim 1, wherein determining the first frequency at which the activity data associated with the set of active users describes the first user further comprises:

determining a uniform resource locator for a web page included in the activity data, the web page including one or more content elements;

mapping the one or more content elements to a cluster;

determining a set of candidate users associated with the cluster;

matching the set of candidate users with the one or more content elements to determine a set of matching scores for the set of candidate users;

determining a top candidate user from the set of candidate users based at least in part on the set of matching scores, the top candidate user having a highest matching score; and

determining whether the first user is the top candidate user.

4. The method of claim 1 further comprising:

parsing the activity data associated with the set of active users that describes the first user;

determining a category for the first user based at least in part on the parsing of the activity data, the category describing one or more occupations of the first user; and

storing the activity data and category data describing the category for the first user in a database.

5. The method of claim 4 further comprising determining the second user based at least in part on the category, the second user having one of the one or more occupations described by the category.

6. The method of claim 1 further comprising:

determining an influence score bucket range; and

determining the second user that has the second influence score based at least in part on the influence score bucket range, the second influence score included in the influence score bucket range.

7. The method of claim 1, wherein the activity data includes a search log, a page view history, a video view history, an audio access history and email content record.

8. A computer program product comprising a non-transitory computer usable medium including a computer readable program, wherein the computer readable program when executed on a computer causes the computer to:

determine whether an influence profile for a first user includes a first influence score on a social network;

responsive to an absence of the first influence score for the first user, determine activity data associated with a set of active users, the activity data including activity of the set of active users from one or more data sources independent of the social network;

determine a first frequency at which the activity data associated with the set of active users describes the first user and a second frequency at which the activity data associated with the set of active users describes a second user, the second user having an influence profile including a second influence score on the social network;

determine a first activity score for the first user based at least in part on the first frequency at which the activity data associated with the set of active users describes the first user, the first activity score for the first user indicating a first interest of the set of active users in the first user;

determine a second activity score for the second user based at least in part on the second frequency at which the activity data associated with set of active users describes the second user, the second activity score for the second user indicating a second interest of the set of active users in the second user;

determine an influence ratio using the first activity score for the first user and the second activity score for the second user, the influence ratio corresponding to a ratio of the first influence score for the first user to the second influence score for the second user; and

determine the first influence score for the first user based at least in part on the influence ratio and the second influence score for the second user.

9. The computer program product of claim 8, wherein determining the first frequency at which the activity data associated with the set of active users describes the first user further comprises:

determining a search query included in the activity data and a set of candidate users based at least in part on the search query;

matching the set of candidate users with one or more web page content elements associated with the search query to determine a set of matching scores for the set of candidate users;

determining a top candidate user from the set of candidate users based at least in part on the set of matching scores, the top candidate user having a highest matching score; and

determining whether the first user is the top candidate user.

10. The computer program product of claim 8, wherein determining the first frequency at which the activity data associated with the set of active users describes the first user further comprises:

determining a uniform resource locator for a web page included in the activity data, the web page including one or more content elements;

mapping the one or more content elements to a cluster;

determining a set of candidate users associated with the cluster;

matching the set of candidate users with the one or more content elements to determine a set of matching scores for the set of candidate users;

determining a top candidate user from the set of candidate users based at least in part on the set of matching scores, the top candidate user having a highest matching score; and

determining whether the first user is the top candidate user.

11. The computer program product of claim 8, wherein the computer readable program when executed on the computer causes the computer to also:

parse the activity data associated with the set of active users that describes the first user;

determine a category for the first user based at least in part on the parsing of the activity data, the category describing one or more occupations of the first user; and

store the activity data and category data describing the category for the first user in a database.

12. The computer program product of claim 11, wherein the computer readable program when executed on the computer causes the computer to also determine the second user based at least in part on the category, the second user having one of the one or more occupations described by the category.

13. The computer program product of claim 8, wherein the computer readable program when executed on the computer causes the computer to also:

determine an influence score bucket range; and

determine the second user that has the second influence score based at least in part on the influence score bucket range, the second influence score included in the influence score bucket range.

14. The computer program product of claim 8, wherein the activity data includes a search log, a page view history, a video view history, an audio access history and email content record.

15. A system comprising:

one or more processors;

an association module stored on a memory and executable by the one or more processors for determining whether an influence profile for a first user includes a first influence score on a social network;

a controller stored on the memory and executable by the one or more processors for determining activity data associated with a set of active users responsive to an absence of the first influence score for the first user, the activity data including activity of the set of active users from one or more data sources independent of the social network;

a determination module stored on the memory and executable by the one or more processors, the determination module communicatively coupled to the controller for determining a first frequency at which the activity data associated with the set of active users describes the first user and a second frequency at which the activity data associated with the set of active users describes a second user, the second user having an influence profile including a second influence score on the social network, the determination module determining a first activity score for the first user based at least in part on the first frequency at which the activity data associated with the set of active users describes the first user, the first activity score for the first user indicating a first interest of the set of active users in the first user and a second activity score for the second user based at least in part on the second frequency at which the activity data associated with the set of active users describes the second user, the second activity score for the second user indicating a second interest of the set of active users in the second user;

a ratio module stored on the memory and executable by the one or more processors, the ratio module communicatively coupled to the determination module for determining an influence ratio using the first activity score for the first user and the second activity score for the second user, the influence ratio corresponding to a ratio of the first influence score for the first user to the second influence score for the second user; and

a score module stored on the memory and executable by the one or more processors, the score module communicatively coupled to the ratio module for determining the first influence score for the first user based at least in part on the influence ratio and the second influence score for the second user.

16. The system of claim 15, wherein determining the first frequency at which the activity data associated with the set of active users describes the first user further comprises:

determining a search query included in the activity data and a set of candidate users based at least in part on the search query;

matching the set of candidate users with one or more web page content elements associated with the search query to determine a set of matching scores for the set of candidate users;

determining a top candidate user from the set of candidate users based at least in part on the set of matching scores, the top candidate user having a highest matching score; and

determining whether the first user is the top candidate user.

17. The system of claim 15, wherein determining the first frequency at which the activity data associated with the set of active users describes the first user further comprises:

determining a uniform resource locator for a web page included in the activity data, the web page including one or more content elements;

mapping the one or more content elements to a cluster;

determining a set of candidate users associated with the cluster;

matching the set of candidate users with the one or more content elements to determine a set of matching scores for the set of candidate users;

determining a top candidate user from the set of candidate users based at least in part on the set of matching scores, the top candidate user having a highest matching score; and

determining whether the first user is the top candidate user.

18. The system of claim 15, further comprising a classification module for parsing the activity data associated with the set of active users that describes the first user, determining a category for the first user based at least in part on the parsing of the activity data, the category describing one or more occupations of the first user and storing the activity data and category data describing the category for the first user in a database.

19. The system of claim 18, wherein the association module further determines the second user based at least in part on the category, the second user having one of the one or more occupations described by the category.

20. The system of claim 15, wherein the association module further determines an influence score bucket range and determines the second user that has the second influence score based at least in part on the influence score bucket range, the second influence score included in the influence score bucket range.