Text-based discourse analysis and management

11620,456

16/859004

April 27, 2020

Systems and methods of the invention determine evasiveness of postings and manage chat sessions accordingly. In embodiments, a method includes accessing a real-time text-based discourse session comprised of multiple text-based posts published by participants, the posts including a question from an author and responses from at least one respondent; determining relationships between words in the text-based discourse session utilizing corpus linguistics analysis; determining a frequency of the responses of the at least one respondent over time; determining an evasiveness score for each of the responses based on natural language processing of the responses, wherein each of the evasiveness scores indicate a level of relevance of a response with respect to the question; determining rankings for each of the responses based on the determined relationships of words, the frequency of the responses, and the evasiveness scores; and determining a display order for the responses based on the rankings of the responses.

INTERNATIONAL BUSINESS MACHINES CORPORATION (Armonk, NY, US)

INTERNATIONAL BUSINESS MACHINES CORPORATION (Armonk, NY, US)

1. A method, comprising: accessing, by a computing device, a real-time text-based discourse session comprised of multiple text-based posts published by participants, the posts including a question from an author and responses from at least one respondent; determining, by the computing device, relationships between words in the real-time text-based discourse session utilizing corpus linguistics analysis; determining, by the computing device, a frequency of the responses of the at least one respondent over time; determining, by the computing device, an evasiveness score for each of the responses based on natural language processing of the responses, wherein each of the evasiveness scores indicate a level of relevance of a response with respect to the question; determining, by the computing device, evasiveness rankings for each of the responses based on the determined relationships between words, the frequency of the responses, and the evasiveness scores; and determining, by the computing device, a display order for the responses based on the evasiveness rankings of the responses.

2. The method of claim 1 , further comprising automatically changing, by the computing device, an order of the responses as they appear in the real-time text-based discourse session from an original display order to the determined display order, wherein the determined display order is different from the original display order.

3. The method of claim 1 , further comprising: determining, by the computing device, that one of the evasiveness scores of one of the responses meets a predetermined threshold value; and generating, by the computing device, an alert based on the one of the evasiveness scores meeting the predetermined threshold value.

4. The method of claim 3 , wherein the generating the alert comprises generating an alert to the author indicating that the one of the responses may be evasive.

5. The method of claim 3 , wherein the generating the alert comprises generating an alert to the at least one respondent indicating that the one of the responses may be evasive.

6. The method of claim 1 , further comprising inserting, by the computing device, at least one indicator of evasiveness in the real-time text-based discourse session based on the evasiveness scores.

7. The method of claim 1 , further comprising assigning, by the computing device, a participant score to one of the participants based on aggregated evasiveness scores of the participant over time.

8. The method of claim 7 , wherein the real-time text-based discourse session is associated with a first community, the method further comprising redirecting, by the computing device, the one of the participants to a second community based on the score.

9. The method of claim 1 , wherein the at least one respondent comprises a chatbot, the method further comprising: determining, by the computing device, that aggregated evasiveness scores of the chatbot over a period of time meet a predetermined threshold value; and providing, by the computing device, feedback to a user based on the aggregated evasiveness scores of the chatbot meeting the predetermined threshold value.

10. The method of claim 1 , further comprising assigning, by the computing device, a category of participant to the at least one respondent based on the evasiveness scores of the responses, wherein the category is one of a plurality of categories associated with different types of access with respect to the real-time text-based discourse session.

11. The method of claim 1 , further comprising: determining, by the computing device, that an increase in evasiveness scores associated with a participant over time meets a threshold increase; and sending, by the computing device, an alert based on the increase in evasiveness scores meeting the threshold increase.

12. The method of claim 1 , further comprising: determining, by the computing device, that a potential sensitive reason exists for at least one of the evasiveness scores of a respondent based on context analysis of user data of the respondent and predetermined sensitivity rules; and alerting, by the computing device, the author of the potential sensitive reason.

13. The method of claim 1 , further comprising: determining, by the computing device, if an evasiveness score of one of the responses of the at least one respondent is malicious or non-malicious based on a sentiment analysis of the response; and at least one selected from the group consisting of: issuing, by the computing device, guidance to one of the participants based on the determining the response is malicious or non-malicious; and filtering, by the computing device, additional responses of the at least one respondent based on the determining the response is malicious or non-malicious.

14. The method of claim 1 , wherein the computing device includes software provided as a service in a cloud environment.

15. A computer program product comprising one or more computer readable storage media having program instructions collectively stored on the one or more computer readable storage media, the program instructions executable to: access a real-time text-based discourse session comprised of multiple text-based posts published by participants, the posts including a question from an author and responses from at least one respondent; determine relationships between words in the real-time text-based discourse session utilizing corpus linguistics analysis; determine a frequency of the responses of the at least one respondent over time; determine an evasiveness score for each of the responses based on natural language processing of the responses, wherein each of the evasiveness scores indicate a level of relevance of a response with respect to the question; determine evasiveness rankings for each of the responses based on the determined relationships between words, the frequency of the responses, and the evasiveness scores; initiate a change in a display order of the responses based on the evasiveness rankings of the responses; and assign a participant score to one of the participants based on aggregated evasiveness scores of the participant over time.

16. The computer program product of claim 15 , wherein the program instructions are further executable to: determine that one of the evasiveness scores of one of the responses meets a predetermined threshold value; and perform at least one selected from the group consisting of: generate an alert based on the one of the evasiveness scores meeting the predetermined threshold value; and insert at least one indicator of evasiveness in the real-time text-based discourse session based on the one of the evasiveness scores meeting the predetermined threshold value.

17. The computer program product of claim 15 , wherein the program instructions are further executable to: assign permissions or access authorizations to the one of the participants based on the aggregate evasiveness scores.

18. A system comprising: a processor, a computer readable memory, one or more computer readable storage media, and program instructions collectively stored on the one or more computer readable storage media, the program instructions executable to: access a real-time text-based discourse session comprised of multiple text-based posts published by participants, the posts including a question from an author and responses from at least one respondent; determine relationships between words in the real-time text-based discourse session utilizing corpus linguistics analysis; determine a frequency of the responses of the at least one respondent over time; determine an evasiveness score for each of the responses based on natural language processing of the responses, wherein each of the evasiveness scores indicate a level of relevance of a response with respect to the question; determine evasiveness rankings for each of the responses, wherein the evasiveness rankings comprise outputs of a model based on an input to the model of the determined relationships between words, the frequency of the responses, and the evasiveness scores; and determine a display order for the responses based on the evasiveness rankings of the responses.

19. The system of claim 18 , wherein the program instructions are further executable to: determine that one of the evasiveness scores of one of the responses meets a predetermined threshold value; and perform at least one selected from the group consisting of: generate an alert based on the one of the evasiveness scores meeting the predetermined threshold value; and insert at least one indicator of evasiveness in the real-time text-based discourse session based on the one of the evasiveness scores meeting the predetermined threshold value.

20. The system of claim 18 , wherein the program instructions are further executable to assign permissions or access authorizations to one of the participants based on aggregate evasiveness scores associated with the one of the participants over time.

9478145 October 2016 Chavez
9740769 August 2017 Bhagwat
10061848 August 2018 Basu
10223442 March 2019 Starr
10831989 November 2020 Takahashi
20040117352 June 2004 Schabes
20100174813 July 2010 Hildreth et al.
20130017524 January 2013 Barborak
20130046531 February 2013 Chandramouli et al.
20140214489 July 2014 Bedi
20150254566 September 2015 Chandramouli et al.
20160226813 August 2016 Cecchi et al.
20180025075 January 2018 Beller
20190108701 April 2019 Kerning et al.
20190260694 August 2019 Londhe
20190318220 October 2019 Beller
20190325360 October 2019 Lam et al.
2019070790 April 2019

Derrick et al., “Detecting Deceptive Chat-Based Communication Using Typing Behavior and Message Cues”, ACM Trans. Manage. Inf. Syst. 4, 2, Article 9, Aug. 2013, 21 pages. cited by applicant
Zhou et al., “A Statistical Language Modeling Approach to Online Deception Detection,” in IEEE Transactions on Knowledge and Data Engineering, vol. 20, No. 8, pp. 1077-1081, Aug. 2008, 5 pages. cited by applicant
Zhou et. al, “An exploratory study into deception detection in text-based computer-mediated communication,” 36th Annual Hawaii International Conference on System Sciences, Proceedings of the, Big Island, HI, USA, 2003, 10 pages. cited by applicant
Rubin, On deception and deception detection: content analysis of computer-mediated stated beliefs. In Proceedings of the 73rd ASIS&T Annual Meeting on Navigating Streams in an Information Ecosystem (ASIS&T '10), vol. 47. American Society for Information Science, Silver Springs, MD, Article 32, 2010, 10 pages. cited by applicant
Mell et al., “The NIST Definition of Cloud Computing”, NIST, Special Publication 800-145, Sep. 2011, 7 pages. cited by applicant
Masip et al., “Learning to Detect Deception from Evasive Answers and Inconsistencies across Repeated Interviews: A Study with Lay Respondents and Police Officers”, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5758596/, Frontiers in Psychology, Jan. 4, 2018, 12 pages. cited by applicant
Wang et al., “Predicting Thread Discourse Structure over TechnicalWeb Forums”, Proceedings of the 2011 Conference on Empirical Methods in Natural Language Processing, pp. 13-25, Edinburgh, Scotland, UK, Jul. 27-31, 2011, 13 pages. cited by applicant
Masip et al., “Strategic Interviewing to Detect Deception: Cues to Deception across Repeated Interviews”, https://www.frontiersin.org/articles/10.3389/fpsyg.2016.01702/full, Frontiers in Psychology, Nov. 1, 2016, 41 pages. cited by applicant

edspgr.11620456