Analyzing distributed group discussions

• US 9,386,107 B1
• Filed: 03/06/2013
• Issued: 07/05/2016
• Est. Priority Date: 03/06/2013
• Status: Active Grant

• Current Assignee: Deloitte Development LLC
• Original Assignee: Blab Incorporated
• Inventors: Snelling, David Aaron, Browning, Randolph, Papp, Stefan Nicholas, Zimmermann, Bernhard G., Downs, Oliver B.
• Primary Examiner(s): Blair, Douglas
• Application Number:
  US13/787,545
• Time in Patent Office: 1,217 Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/2465 Query processing support for facilitating data mining operations in structured databases
• CPC Class Codes: G06Q 50/01 Social networking
• CPC Class Codes: H04L 51/02 using automatic reactions or user delegation, e.g. automatic replies or chatbot-generated messages
• CPC Class Codes: H04L 51/04 Real-time or near real-time messaging, e.g. instant messaging [IM]
• CPC Class Codes: H04L 51/216 Handling conversation history, e.g. grouping of messages in sessions or threads
• CPC Class Codes: H04L 51/52 for supporting social networking services

Abstract

Techniques are described for analyzing user-supplied information, including in at least some situations to predict future aspects of additional related information that will be supplied by users. The user-supplied information that is analyzed may, for example, include distributed group discussions that involve numerous users and occur via user comments made to one or more social networking sites and/or other computer-accessible sites. The analysis of user-supplied information may, for example, include determining particular topics that are of interest for a specified category during one or more periods of time, quantifying an amount of user interest in particular topics and the category during the period of time, predicting future amounts of user interest in the particular topics and the category during one or more future period of times, and taking one or more further actions based on the predicted information.

Claims

1. A computer-implemented method comprising:

obtaining, by one or more configured computing systems and via electronic interactions over one or more computer networks, information from social networking sites including a plurality of textual comments that are supplied by human users to the social networking sites over multiple prior time periods;

determining, by the one or more configured computing systems, multiple topics associated with a specified content category based on contents of the plurality of textual comments, by;

analyzing the contents of the plurality of textual comments to identify a plurality of topics that are mentioned in the contents;

determining, for each of the identified topics, a quantity of the plurality of textual comments having contents that mention the identified topic and that match one or more keywords for the specified content category; and

selecting a subset of the identified topics to be the determined multiple topics associated with the specified content category, the selecting including excluding one or more of the identified topics whose determined quantity is less than a minimum threshold or more than a maximum threshold;

analyzing, by the one or more configured computing systems, and for each of the multiple prior time periods, multiple textual comments from the plurality that are supplied by human users during the prior time period to identify a quantity of the multiple textual comments having contents that mention an indicated one of the determined multiple topics;

matching, by the one or more configured computing systems, the identified quantities for the multiple prior time periods to a first portion of a defined prediction template having information representing changes in quantity over time;

predicting, by the one or more configured computing systems, expected quantities of future textual comments that will be supplied during multiple future time periods and will be associated with the indicated one topic, the predicting being based at least in part on a second portion of the defined prediction template that is distinct from the first portion; and

providing, by the one or more configured computing systems and via additional electronic interactions over the one or more computer networks, information to one or more recipients that includes indications of the predicted expected quantities of the future textual comments.

2. The method of claim 1 further comprising:

determining, by the one or more configured computing systems and based at least in part on the predicted expected quantities of the future textual comments that will be associated with the indicated one topic, one or more types of information to make available to users corresponding to the indicated one topic and the specified content category; and

supplying to one or more of the social networking sites, by the one or more configured computing systems, one or more additional textual comments that have contents corresponding to the determined one or more types of information.

3. The method of claim 1 further comprising generating, by the one or more configured computing systems, a first histogram to represent the identified quantities for the multiple prior time periods, and wherein the defined prediction template includes a second histogram for a length of time longer than the multiple prior time periods that includes information based on prior actual quantities of textual comments received, and wherein the matching of the identified quantities for the multiple prior time periods to the first portion of the defined prediction template includes matching the first histogram to a subset of the second histogram.

4. The method of claim 1 further comprising:

predicting, by the one or more configured computing systems, a second group of expected quantities of future textual comments that will be supplied during the multiple future time periods and will be associated with the indicated one topic, the predicting of the second group being based at least in part on a weighted combination of the identified quantities for the multiple prior time periods;

determining, by the one or more configured computing systems at a time after the multiple future time periods, that the expected quantities for the predicted second group match actual quantities of textual comments associated with the indicated one topic that were supplied during the multiple future time periods; and

generating, by the one or more configured computing systems in response to the determining that the expected quantities for the predicted second group match the actual quantities, a second prediction template for use in predicting additional expected quantities of additional future textual comments, the generated second prediction template being based at least in part on the identified quantities for the multiple prior time periods and on the actual quantities for the multiple future time periods.

5. A computer-implemented method comprising:

obtaining, by one or more configured computing systems and via electronic interactions over one or more computer networks, information from one or more network-accessible sites including a plurality of comments that are supplied by human users to the one or more network-accessible sites over multiple prior time periods and that are related to multiple topics;

analyzing, by the one or more configured computing systems, and for each of the multiple prior time periods, multiple textual comments from the plurality that are for the prior time period to identify a subset of the multiple textual comments whose contents are associated with a specified content category;

determining, by the one or more configured computing systems, and for each of the multiple prior time periods, a quantity of the identified textual comments for the prior time period that are associated with an indicated topic from the multiple topics;

predicting, by the one or more configured computing systems, and for each of multiple future time periods, an expected quantity of future textual comments associated with the indicated topic that will be supplied by human users during the future time period, the predicting being based at least in part on the determined quantities for the multiple prior time periods; and

providing, by the one or more configured computing systems and via additional electronic interactions over the one or more computer networks, information to one or more recipients that includes, for each of one or more of the multiple future time periods, an indication of the predicted expected quantity of future textual comments for the future time period.

6. The method of claim 5 wherein the predicting of expected quantities of future textual comments for the multiple future time periods includes generating trend information based on the determined quantities of identified textual comments for at least some of the prior time periods, and determining the predicted expected quantities for at least some of the multiple future time periods based on the generated trend information.

7. The method of claim 6 wherein the generating of the trend information includes calculating a trend line based on a weighted combination of the determined quantities of identified textual comments for the at least some prior time periods.

8. The method of claim 5 wherein the predicting of expected quantities of future textual comments for the multiple future time periods includes matching the determined quantities of identified textual comments for at least some of the prior time periods to a first portion of a defined template representing changes in quantity over time, and determining the predicted expected quantities for at least some of the multiple future time periods based on information from a second portion of the defined template.

9. The method of claim 8 further comprising generating, by the one or more configured computing systems, a first histogram to represent the determined quantities of identified textual comments for the at least some prior time periods, and wherein the defined template includes a second histogram based on prior actual quantities of textual comments received, and wherein the matching of the determined quantities for the at least some prior time periods to the first portion of the defined template includes matching the first histogram to a subset of the second histogram.

10. The method of claim 8 further comprising, before the analyzing of the multiple textual comments for each of the multiple prior time periods, automatically generating, by the one or more configured computing systems, the defined template based at least in part on additional textual comments received for multiple additional time periods that occurred before the multiple prior time periods.

11. The method of claim 10 further comprising, before the analyzing of the multiple textual comments for each of the multiple prior time periods:

using, by the one or more configured computing systems and for each of multiple further times that occurred before the multiple prior time periods, the defined template to provide additional predictions of expected future quantities of textual comments;

tracking, by the one or more configured computing systems and based on the provided additional predictions for the multiple further times, an accuracy rate of the defined template in providing accurate predictions; and

determining, by the one or more configured computing systems, to continue using the defined template based at least in part on the tracked accuracy rate,

and wherein the matching of the determined quantities of identified textual comments for at least some of the prior time periods to the first portion of the defined template is based at least in part on the determining to continue using the defined template.

12. The method of claim 10 further comprising:

using, by the one or more configured computing systems and for each of multiple further times, the generated defined template to provide additional predictions of expected future quantities of textual comments;

tracking, by the one or more configured computing systems and based on the provided additional predictions for the multiple further times, an accuracy rate of the generated defined template in providing accurate predictions; and

determining, by the one or more configured computing systems, to remove the generated defined template from further use based at least in part on the tracked accuracy rate.

13. The method of claim 10 further comprising, before the analyzing of the multiple textual comments for each of the multiple prior time periods:

using, by the one or more configured computing systems and for each of multiple further times that occurred before the multiple prior time periods, the generated defined template to provide additional predictions of expected future quantities of textual comments;

determining, based at least in part on success from the using of the generated defined template for the multiple further times, to promote further use of the generated defined template,

and wherein the matching of the determined quantities of identified textual comments for at least some of the prior time periods to the first portion of the defined template is based at least in part on the determining to promote the further use of the generated defined template.

14. The method of claim 5 further comprising generating, by the one or more configured computing systems, a defined template representing changes in quantity over time based at least in part on the determined quantities of identified textual comments for the multiple prior time periods, to enable predictions of additional future quantities of textual comments when future conditions match at least a portion of the defined template.

15. The method of claim 14 wherein the generating of the defined template is further based in part on the predicted expected quantities for at least some of the multiple future time periods, and includes verifying, at a time after the at least some future time periods, that the predicted expected quantities for the at least some future time periods match actual quantities of textual comments associated with the indicated topic that were supplied during the at least some future time periods.

16. The method of claim 5 wherein the one or more configured computing systems are part of a group discussion prediction service, and wherein the specified content category is indicated by a client of the group discussion prediction service that is at least one of the recipients.

17. The method of claim 5 further comprising:

determining, by the one or more configured computing systems and based at least in part on the predicted expected quantities for the one or more future time periods, one or more types of information to make available corresponding to at least one of the indicated topic and the specified content category; and

supplying, by the one or more configured computing systems, one or more additional textual comments that are available to the human users and that have contents corresponding to the determined one or more types of information.

18. The method of claim 17 wherein the multiple textual comments for each of the multiple prior time periods are part of user posts to one or more social networking services by the human users, and wherein the supplying of the one or more additional textual comments includes supplying the additional textual comments to at least one of the social networking services as additional posts.

19. The method of claim 5 wherein the plurality of comments are supplied by the human users to the one or more network-accessible sites from a plurality of geographical locations, and wherein the method further comprises:

determining, based at least in part on a first subset of the plurality of comments that are associated with a first geographical location of the plurality of geographical locations, an increase or a decrease in a first quantity of comments in the first subset relative to a second quantity of comments in a second subset of the plurality of comments that are associated with one or more other second geographical locations of the plurality of geographical locations; and

providing one or more indications of the first geographical location and of the determined increase or decrease.

20. The method of claim 5 wherein the plurality of comments are supplied by a plurality of human users, and wherein the method further comprises:

determining, based at least in part on a first subset of the plurality of comments that are supplied by a first author user of the plurality of human users, an increase or a decrease in a first quantity of comments in the first subset relative to a second quantity of comments in a second subset of the plurality of comments that are supplied by one or more other second author users of the plurality of human users; and

providing one or more indications of the first author user and of the determined increase or decrease.

21. The method of claim 5 wherein the plurality of comments are supplied by the human users to a plurality of information sources, and wherein the method further comprises:

determining, based at least in part on a first subset of the plurality of comments that are associated with a first information source of the plurality of information sources, an increase or a decrease in a first quantity of comments in the first subset relative to a second quantity of comments in a second subset of the plurality of comments that are associated with one or more other second information sources of the plurality of information sources; and

providing one or more indications of the first information source and of the determined increase or decrease.

22. The method of claim 5 wherein the plurality of comments are supplied by a plurality of human users, and wherein the method further comprises:

performing an influence analysis of the plurality of comments to identify a first author user of the plurality of human users who has greater influence in affecting comments of other users than one or more other second author users of the plurality of human users; and

providing one or more indications of the first author user and the performed influence analysis.

23. The method of claim 5 further comprising:

determining, based at least in part on the multiple textual comments for a first time period of the multiple prior time periods, an increase or a decrease in a positive sentiment or a negative sentiment relative to the multiple textual comments for one or more other second time periods of the multiple prior time periods; and

providing one or more indications of the determined increase or decrease in the positive sentiment or the negative sentiment.

24. A non-transitory computer-readable medium having stored contents that cause a computing system to perform a method, the method comprising:

obtaining, by the computing system and via electronic interactions over one or more computer networks, information from a plurality of online information sources including a plurality of user-supplied content items that are associated with multiple time periods and that are supplied to the online information sources;

analyzing, by the computing system, the plurality of user-supplied content items, the analyzing including determining, for each of the multiple time periods, a quantity of the plurality of content items that are associated with the time period and have an indicated attribute;

determining, by the computing system and based on multiple first content items of the plurality of content items that are supplied to a first information source of the plurality of information sources, an increase or a decrease for the first information source relative to one or more other second information sources of the plurality of information sources, wherein the determining of the increase or decrease is of a first quantity of the first content items relative to a second quantity of multiple second content items of the plurality of content items that are supplied to the one or more other second information sources;

predicting, by the computing system and based at least in part on the determined quantities for the multiple time periods, an expected quantity of additional content items having the indicated attribute that will be supplied by users during an additional time period after the multiple time periods; and

providing, by the computing system and via additional electronic interactions over the one or more computer network, information to one or more recipients that includes one or more indications of the predicted expected quantity of additional content items and of the determined increase or decrease for the first information source relative to the one or more other second information sources.

25. The non-transitory computer-readable medium of claim 24 wherein the plurality of content items are comments supplied by a plurality of human users, wherein the plurality of online information sources include one or more social networking services, wherein the indicated attribute is one of a plurality of terms included in contents of the content items of the identified subset, and wherein the stored contents include software instructions that, when executed, further cause the computing system to:

determine, by the computing system and based on multiple third content items of the plurality of content items that are supplied by a first user of the plurality of human users, a further increase or a decrease in a third quantity of the third content items relative to a fourth quantity of multiple fourth content items of the plurality of content items that are supplied by one or more other second users of the plurality of human users; and

provide, by the computing system, an indication of the determined further increase or decrease.

26. The non-transitory computer-readable medium of claim 24 wherein the predicting of the expected quantity of additional content items having the indicated attribute is performed for each of multiple future time periods, wherein the providing of the indication of the predicted expected quantity includes indicating the predicted expected quantities for the multiple future time periods, and wherein the determining of the increase or the decrease for the first information source relative to the one or more other second information sources is further based in part on prior quantities of content items supplied to the first information source and to the one or more other second information sources.

27. The non-transitory computer-readable medium of claim 24 wherein the computer-readable medium is a memory of the computing system, wherein the stored contents are instructions that, when executed, program the computing system to perform the method, and wherein the method further comprises:

determining, by the computing system and based on at least one of the predicted expected quantity of additional content items or the determined increase or decrease for the first information source, one or more types of information to make available to users corresponding to the indicated attribute; and

supplying, by the computing system and to one or more social networking sites, one or more additional content items that have contents corresponding to the determined one or more types of information and to the indicated attribute.

28. A system, comprising:

one or more processors of one or more computing systems; and

one or more memories including stored instructions that, when executed by at least one of the one or more processors, cause the at least one processor to;

obtain, via electronic interactions over one or more computer networks, information from one or more network-accessible sites including a plurality of textual comments that are supplied by multiple users to the one or more network-accessible sites over multiple prior time periods and that are related to multiple topics;

analyze, for each of the multiple prior time periods, multiple textual comments from the plurality that are for the prior time period and identify a subset of the multiple textual comments whose contents are associated with a specified content category;

determine, for each of the multiple prior time periods, an actual quantity of the identified textual comments for the prior time period that are associated with an indicated topic from the multiple topics;

predict, for each of multiple future time periods, an expected quantity of future textual comments associated with the indicated topic that will be supplied by users during the future time period, the predicting being based at least in part on the determined actual quantities for the multiple prior time periods and by using a first defined template representing a first set of changes in quantity over time;

generate, from the determined actual quantities for the multiple prior time periods, a new defined template that includes the determined actual quantities for the multiple prior time periods and that represents a second set of changes in quantity of textual comments over time;

provide, via additional electronic interactions over the one or more computer networks, information to one or more recipients that includes, for each of one or more of the multiple future time periods, an indication of the predicted expected quantity of future textual comments for the future time period; and

use, at a later time after the generating of the new defined template, the new defined template to provide additional predictions of expected future quantities of textual comments for additional future time periods based on additional actual quantities of textual comments received preceding the additional future time periods.

29. The system of claim 28 wherein the predicting of expected quantities of future textual comments for the multiple future time periods includes generating trend information based on the determined actual quantities of identified textual comments for at least some of the prior time periods, and determining the predicted expected quantities for at least some of the multiple future time periods based on the generated trend information.

30. The system of claim 29 wherein the generating of the trend information includes calculating a trend line based on a weighted combination of the determined actual quantities of identified textual comments for the at least some prior time periods.

31. The system of claim 28 wherein the predicting of expected quantities of future textual comments for the multiple future time periods includes matching the determined actual quantities of identified textual comments for at least some of the prior time periods to a first portion of the first defined template, and determining the predicted expected quantities for at least some of the multiple future time periods based on information from a second portion of the first defined template.

32. The system of claim 31 wherein the stored instructions include software instructions that, when executed by the at least one processor, further cause the at least one processor to generate a first histogram to represent the determined actual quantities of identified textual comments for the at least some prior time periods, wherein the first defined template includes a second histogram based on prior actual quantities of textual comments received, and wherein the matching of the determined actual quantities for the at least some prior time periods to the first portion of the first defined template includes matching the first histogram to a subset of the second histogram.

33. The system of claim 28 wherein the generating of the new defined template is further based in part on the predicted expected quantities of future textual comments for at least some of the multiple future time periods.

34. The system of claim 33 wherein the generating of the new defined template further includes verifying, at a time after the at least some future time periods, that the predicted expected quantities for the at least some future time periods match actual quantities of textual comments associated with the indicated topic that were supplied during the at least some future time periods.

35. The system of claim 28 wherein the system is part of a group discussion prediction service, and wherein the specified content category is indicated by a client of the group discussion prediction service.

36. The system of claim 28 wherein the stored instructions include software instructions that, when executed by the at least one processor, further cause the at least one processor to:

determine, based at least in part on the predicted expected quantities for the one or more future time periods, one or more types of information to make available corresponding to at least one of the indicated topic and the specified content category; and

supply one or more additional textual comments that are available to the users and that have contents corresponding to the determined one or more types of information.

37. The system of claim 36 wherein the multiple textual comments for each of the multiple prior time periods are part of user posts to one or more social networking services implemented by the one or more network-accessible sites, and wherein the supplying of the one or more additional textual comments includes supplying the additional textual comments to at least one of the social networking services as additional posts.

38. The system of claim 28 wherein the plurality of textual comments are supplied by human users from a plurality of geographical locations, and wherein the stored instructions include software instructions that, when executed by the at least one processor, further cause the at least one processor to:

determine, based at least in part on a first subset of the plurality of textual comments that are associated with a first geographical location of the plurality of geographical locations, an increase or a decrease in a first quantity of textual comments in the first subset relative to a second quantity of textual comments in a second subset of the plurality of textual comments that are associated with one or more other second geographical locations of the plurality of geographical locations; and

provide one or more indications of the first geographical location and of the determined increase or decrease.

39. The system of claim 28 wherein the plurality of textual comments are supplied by a plurality of human users, and wherein the stored instructions include software instructions that, when executed by the at least one processor, further cause the at least one processor to:

determine, based at least in part on a first subset of the plurality of textual comments that are supplied by a first author user of the plurality of human users, an increase or a decrease in a first quantity of textual comments in the first subset relative to a second quantity of textual comments in a second subset of the plurality of textual comments that are supplied by one or more other second author users of the plurality of human users; and

provide one or more indications of the first author user and of the determined increase or decrease.

40. The system of claim 28 wherein the plurality of textual comments are supplied to a plurality of information sources, and wherein the stored instructions include software instructions that, when executed by the at least one processor, further cause the at least one processor to:

determine, based at least in part on a first subset of the plurality of textual comments that are associated with a first information source of the plurality of information sources, an increase or a decrease in a first quantity of textual comments in the first subset relative to a second quantity of textual comments in a second subset of the plurality of textual comments that are associated with one or more other second information sources of the plurality of information sources; and

provide one or more indications of the first information source and of the determined increase or decrease.

41. The system of claim 28 wherein the plurality of textual comments are supplied by a plurality of human users, and wherein the stored instructions include software instructions that, when executed by the at least one processor, further cause the at least one processor to:

perform an influence analysis of the plurality of textual comments to identify a first author user of the plurality of human users who has greater influence in affecting textual comments of other users than one or more other second author users of the plurality of human users; and

provide one or more indications of the first author user and the performed influence analysis.

42. The system of claim 28 wherein the stored instructions include software instructions that, when executed by the at least one processor, further cause the at least one processor to:

determine, based at least in part on the multiple textual comments for a first time period of the multiple prior time periods, an increase or a decrease in a positive sentiment or a negative sentiment relative to the multiple textual comments for one or more other second time periods of the multiple prior time periods; and

provide one or more indications of the determined increase or decrease in the positive sentiment or the negative sentiment.

43. The system of claim 28 wherein the using of the new defined template to provide additional predictions of expected future quantities of textual comments is performed for each of multiple later times and further includes:

tracking, based on the provided additional predictions for the multiple later times, an accuracy rate of the new defined template in providing accurate predictions; and

determining to continue using the new defined template for further later times based at least in part on the tracked accuracy rate.

44. The system of claim 28 wherein the using of the new defined template to provide additional predictions of expected future quantities of textual comments is performed for each of multiple later times and further includes:

tracking, based on the provided additional predictions for the multiple later times, an accuracy rate of the new defined template in providing accurate predictions; and

determining to remove the new defined template from further use based at least in part on the tracked accuracy rate.

45. The system of claim 28 wherein the stored instructions further cause the at least one processor to, before the using of the new defined template at the later time to provide additional predictions of expected future quantities of textual comments:

using, for each of multiple further times before the later time and after the generating of the new defined template, the new defined template to predict further expected future quantities of textual comments for the further time; and

determining, based at least in part on success from the using of the new defined template for the multiple further times, to promote further use of the new defined template,

and wherein the using of the new defined template at the later time is based at least in part on the determining to promote the further use of the new defined template.