METHOD FOR DETERMINING INTERPERSONAL RELATIONSHIP INFLUENCE INFORMATION USING TEXTUAL CONTENT FROM INTERPERSONAL INTERACTIONS

• US 20130013706A1
• Filed: 08/27/2012
• Published: 01/10/2013
• Est. Priority Date: 07/07/2011
• Status: Abandoned Application

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Gupta, Ajay K., Lotlikar, Rohit M., Rai, Angshu
• Application Number:
  US13/594,963
• Publication Number:
  US 20130013706A1
• Time in Patent Office: Days
• US Class Current: 709/206
• CPC Class Codes: G06Q 10/10 Office automation; Time management

Abstract

A computer-implemented method determines interpersonal relationship influence information between participants in an interaction channel that includes inputting an interaction text response of communication by a participant made in an interaction channel. A type of influence relation of the interaction text response is determined on at least one other participant in the interaction channel based on relational descriptors. An influence score value for the interaction text response is determined, where the influence score value indicates how strongly the interaction text response influences the at least one other participant in the interaction channel. A pair-wise indication of a direction of influence of at least two participants in the interaction channel is determined. Interpersonal relationship influence information is output for the interaction text response for a participant pair in the interaction channel.

Claims

1. A computer-implemented method that determines interpersonal relationship influence information between participants in an interaction channel, said method comprising:

inputting, an interaction text response to a communication by a participant made in an interaction channel;

determining, a type of influence relation said interaction text response has, based on relational descriptors;

determining, an influence score value for said interaction text response;

determining, a pair-wise indication of a direction of influence of at least one participant pair in said interaction channel; and

outputting, for said participant pair in said interaction channel, said interpersonal relationship influence information for said interaction text response comprising said type of influence relation, said influence score value, and said pair-wise indication of a direction of influence.

2. The computer-implemented method according to claim 1, said relation descriptors further comprising keywords, regular expression rules and classification models.

3. The computer-implemented method according to claim 2, said determining of said type of influence relation further comprises:

annotating, said interaction text response based on said keywords and regular expression rules to output a set of annotation flags corresponding to said interaction text response.

4. The computer-implemented method according to claim 2, said determining of said type of influence relation further comprises:

classifying, said interaction text response based on said classification models to output class labels corresponding to said interaction text response.

5. The computer-implemented method according to claim 1, said influence score value representing an extent of influence caused by said interaction text response on one or more participants in said interaction channel.

6. The computer-implemented method according to claim 1, said interpersonal relationship influence information comprises:

said type of influence relation between said participant pair;

said influence score value between said participant pair; and

said pair-wise indication of said direction of influence between said participant pair.

7. The computer-implemented method according to claim 1, said interaction channel comprising at least one of:

e-mail;

a social media channel;

an online bulletin board system;

an online chat room; and

an online forum.

9. The computer-implemented method according to claim 7, said relation descriptors further comprising keywords, regular expression rules and classification models.

10. The computer-implemented method according to claim 9, said annotating outputs said set of annotation flags based on said keywords and regular expression rules.

11. The computer-implemented method according to claim 9, said classifying outputs said class labels based on said classification models.

8. A computer-implemented method that determines interpersonal relationship influence information between participants in an interaction channel, said method comprising:

inputting, an interaction text response to a communication by a participant made in an interaction channel;

annotating, said interaction text response based on relational descriptors to output a set of annotation flags corresponding to said interaction text response;

classifying, said interaction text response based on said relation descriptors to output class labels corresponding to said interaction text response;

determining, a type of influence relation said interaction text response has, based on said set of annotation flags and said class labels;

determining, an influence score value for said interaction text response based on said annotating and said classifying, said influence score value representing an extent of influence caused by said interaction text response on one or more participants in said interaction channel;

determining, a pair-wise indication of a direction of influence of at least one participant pair in said interaction channel; and

outputting, for said participant pair in said interaction channel, interpersonal relationship influence information representing said type of influence relation between said participant pair, said influence score value between said participant pair, and said pair-wise indication of said direction of influence between said participant pair based at least on said type of influence relation, said influence score value, and said pair-wise indication of a direction of influence.

12. The computer-implemented method according to claim 8, said interaction channel comprising one of:

e-mail;

a social media channel;

an online bulletin board system;

an online chat room; and

an online forum.

13. A computer-implemented method that determines interpersonal relationship influence information between participants in an interaction channel, said method comprising:

inputting, an interaction text response to a communication by a participant made in an interaction channel;

determining, a type of influence relation of said interaction text response on at least one other participant in said interaction channel based on relational descriptors;

determining, an influence score value for said interaction text response, said influence score value indicates how strongly said interaction text response influences said at least one other participant in said interaction channel;

determining, a pair-wise indication of a direction of influence of at least one participant pair in said interaction channel; and

outputting, for said participant pair in said interaction channel, interpersonal relationship influence information for said interaction text response based at least on said type of influence relation, said influence score value, and said pair-wise indication of a direction of influence.

14. The computer-implemented method according to claim 13, further comprising:

updating, said relational descriptors based on a particular application requirement.

15. The computer-implemented method according to claim 13, where said type of influence relation further includes information regarding a position in said interaction text response where a particular regular expression rule has been activated.

16. The computer-implemented method according to claim 13, where said type of influence relation further includes class labels that have a class probability value representing a certainty for a class label corresponding to said input interaction text response.

17. The computer-implemented method according to claim 13, where said type of influence relation further comprises a set of annotation flags used as features for classification in an interaction classification system.

18. The computer-implemented method according to claim 13, where said type of influence relation further comprises class labels that cause certain annotation rules to be executed by an interaction annotation system.

19. The computer-implemented method according to claim 13, where said type of influence relation further causes preprocessing of said input interaction text response based upon a set of annotation flags to cause class labels to be output by an interaction classification system.

20. The computer-implemented method according to claim 13, said influence score value comprising one of:

an independent score value for each interaction text response; and

a participant pair score value for a pair of participants based on said interaction text response.