Method and system for identifying a key influencer in social media utilizing topic modeling and social diffusion analysis

• US 8,312,056 B1
• Filed: 09/13/2011
• Issued: 11/13/2012
• Est. Priority Date: 09/13/2011
• Status: Expired due to Fees

• Current Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Original Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Inventors: Peng, Wei, Sun, Tong
• Primary Examiner(s): LEWIS, CHERYL RENEA
• Application Number:
  US13/231,321
• Time in Patent Office: 427 Days
• Field of Search: 707/802, 705/319
• US Class Current: 707/802
• CPC Class Codes: G06F 16/95 Retrieval from the web
• CPC Class Codes: G06Q 50/01 Social networking

Abstract

A system and method for identifying a key influencer in a social media environment for enterprise marketing utilizing topic modeling and social diffusion analysis. A user interest profile can be generated by analyzing historical data stored in a database utilizing. A social graph can be generated and an influence measuring process based on the social graph data can be performed utilizing a static diffusion model and a dynamic diffusion model to calculate a set of key influencers. The dynamic diffusion model considers time stamp information to assess an impact of each user communication on the growth of a conversation within a time period. The key influencer can be identified in a specific topic area and a number of total users that can be reached via the influencer within a specific time window can be predicted.

Claims

1. A method for identifying a key influencer in a social media, said method comprising:

compiling a user interest profile by analyzing a historical message stored in a database utilizing a topic modeling approach to thereafter generate social graph data;

performing an influence measuring process based on said social graph data utilizing an influence measuring module and a data intensive distributed application module to identify a set of key influencers in a particular topic area by considering time stamp information; and

assessing an impact of user communication with respect to growth of a conversation within a predefined time interval in order to predict a message propagation speed and coverage in a social network and permit efficient targeting of said message to said key influencer of a targeted community.

2. The method of claim 1 wherein said topic modeling approach comprises a Latent Dirichlet Allocation modeling process.

3. The method of claim 2 wherein said topic modeling approach further comprises:

extracting a topic from a raw text and a plurality of social contents posted by said user and/or a hashtag so that said user and said hashtag is related to a set of messages; and

performing a spammer filtering and text preprocessing approach before modeling said user interest topic.

4. The method of claim 1 further comprising generating said social graph utilizing said user as a node and a social relationship between said users as an edge wherein said edge weight is defined based on a strength of relationship.

5. The method of claim 1 wherein said influence measuring process comprises a static diffusion model.

6. The method of claim 5 wherein said static diffusion model comprises at least one of the following types of model:

a betweenness centrality analysis; and

a K-Shell decomposition process.

7. The method of claim 6 wherein said betweenness centrality analysis further comprises:

measuring centrality of a node within a graph by said betweenness centrality analysis in order to emphasize strength of a weak tie between at least two clustered communities having an impact on coverage of a marketing message diffusion process.

8. The method of claim 1 wherein said K-Shell decomposition process further comprises considering an individual in a highest K-shell as said influential user.

9. The method of claim 1 wherein said influence measuring process comprises a dynamic diffusion model.

10. The method of claim 9 wherein said dynamic diffusion model comprises at least one of the following types of model:

a diffusion graph construction process; and

a continuous-time Markov process.

11. The method of claim 10 wherein said diffusion graph construction process further comprises:

generating a propagation sequence based on a hashtag and a uniform resource locator and sequencing a group of users based on a time of post for a given hashtag and uniform resource locator;

categorizing said propagation sequence into a plurality of topics utilizing said topic modeling approach wherein said propagation sequence in each topic constructs a topic-dependent diffusion graph that naturally considers a topic-dependent user interest.

12. The method of claim 10 wherein said continuous-time Markov process further comprises:

denoting a transition probability from one state to another state over a time period and extracting a message propagation state utilizing said time stamp information;

computing a user rank by calculating a transition rate matrix based on an extracted state in order to thereafter generate a transition matrix and a steady state distribution of said transition matrix;

estimating a mean first passage matrix and a mean first passage time matrix in order to calculate said user rank; and

performing a propagation estimation by estimating a transition probability matrix to identify said key influencer.

13. The method of claim 1 further comprising accelerating influence measuring process utilizing said data intensive distributed application module to calculate said key influencer within said specific time window.

14. A system for identifying a key influencer in a social media, said system comprising:

a processor; and

a computer-usable medium embodying computer code, said computer-usable medium being coupled to said data bus, said computer program code comprising instructions executable by said processor and configured for;

compiling a user interest profile by analyzing a historical message stored in a database utilizing a topic modeling approach to thereafter generate social graph data;

performing an influence measuring process based on said social graph data utilizing an influence measuring module and a data intensive distributed application module to identify a set of key influencers in a particular topic area by considering time stamp information; and

assessing an impact of user communication with respect to growth of a conversation within a predefined time interval in order to predict a message propagation speed and coverage in a social network and permit efficient targeting of said message to said key influencer of a targeted community.

15. The system of claim 14 wherein said topic modeling approach comprises a Latent Dirichlet Allocation modeling process.

16. The system of claim 14 wherein said instructions are further configured for generating said social graph utilizing said user as a node and a social relationship between said users as an edge wherein said edge weight is defined based on a strength of relationship.

17. The system of claim 14 wherein said influence measuring process comprises a static diffusion model.

18. The system of claim 14 wherein said influence measuring process comprises a dynamic diffusion model.

19. The system of claim 14 wherein said instructions are further configured for accelerating influence measuring process utilizing said data intensive distributed application module to calculate said key influencer within said specific time window.

20. A processor-readable medium storing code representing instructions to cause a processor to perform a process for identifying a key influencer in a social media, said code comprising of code to:

compile a user interest profile by analyzing a historical message stored in a database utilizing a topic modeling approach to thereafter generate social graph data;

perform an influence measuring process based on said social graph data utilizing an influence measuring module and a data intensive distributed application module to identify a set of key influencers in a particular topic area by considering time stamp information; and

assess an impact of user communication with respect to growth of a conversation within a predefined time interval in order to predict a message propagation speed and coverage in a social network and permit efficient targeting of said message to said key influencer of a targeted community.