System and method for scoring electronic messages

• US 7,600,017 B2
• Filed: 01/10/2007
• Issued: 10/06/2009
• Est. Priority Date: 10/11/2000
• Status: Expired due to Term

• Current Assignee: Buzzmetrics, Ltd. (Nielsen Holdings plc)
• Original Assignee: Buzzmetrics, Ltd. (Nielsen Holdings plc)
• Inventors: Kodey, Robert, Holtzman, David, Pool, David
• Primary Examiner(s): Bruckart; Benjamin R
• Application Number:
  US11/651,661
• Publication Number:
  US 20070124432A1
• Time in Patent Office: 1,000 Days
• Field of Search: 709217-219, 709/224, 704/9
• US Class Current: 709/224
• CPC Class Codes: G06F 21/31 User authentication
• CPC Class Codes: G06F 2221/2117 User registration
• CPC Class Codes: G06F 40/258 Heading extraction; Automatic titling; Numbering
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 40/00 Finance; Insurance; Tax strategies; Processing of corporate or income taxes
• CPC Class Codes: G06Q 40/06 Asset management; Financial planning or analysis
• CPC Class Codes: G06Q 50/01 Social networking
• CPC Class Codes: H04L 12/1831 Tracking arrangements for later retrieval, e.g. recording contents, participants activities or behavior, network status
• CPC Class Codes: H04L 51/216 Handling conversation history, e.g. grouping of messages in sessions or threads
• CPC Class Codes: H04L 63/0407 wherein the identity of one or more communicating identities is hidden cryptographic mechanisms or cryptographic arrangements for anonymous credentials or for identity based cryptographic systems H04L9/00
• CPC Class Codes: H04L 65/1101 Session protocols
• CPC Class Codes: H04L 9/40 Network security protocols
• CPC Class Codes: Y10S 707/99937 Sorting

Abstract

A system and method for collecting and analyzing electronic discussion messages to categorize the message communications and the identify trends and patterns in pre-determined markets. The system comprises an electronic data discussion system wherein electronic messages are collected and analyzed according to characteristics and data inherent in the messages. The system further comprises a data store for storing the message information and results of any analyses performed. Objective data is collected by the system for use in analyzing the electronic discussion data against real-world events to facilitate trend analysis and event forecasting based on the volume, nature and content of messages posted to electronic discussion forums.

Claims

1. A system for processing message traffic from a plurality of electronic discussion forums, comprising:

a computer implemented message collector for collecting messages from the plurality of electronic discussion forums;

a computer including processing logic to process the messages based on a topic in order to track a plurality of pseudonyms, wherein the processing logic computes at least one numerical relevance score for the collected messages posted by the plurality of pseudonyms, wherein the numerical relevance score rates the relevance of the collected messages to the topic on a numerical scale, and wherein the numerical relevance score is determined based on a set of predefined rules, the set of predefined rules having a plurality of conditions defining information relevant to the topic and having an associated weighting to indicate the strength a particular condition should have in determining the relevance of the messages with respect to the topic; and

a computer implemented report generator for generating a report on the topic using results from the processing logic, wherein the report at least indirectly reflects at least the at least one numerical relevance score for the collected messages posted by the plurality of pseudonyms;

wherein the processing logic is further adapted to compute a buzz score based on a set of the collected messages for the topic, and wherein the buzz score is a function of the at least one numerical relevance score for messages in the set of the collected messages, and the report at least indirectly reflects the buzz score.

2. The system of claim 1, wherein the buzz score is a function of numerical influence scores of each of the posting pseudonyms that respectively posted the set of the collected messages, and wherein the influence score rates the influence of the posting pseudonym on a numerical scale.

3. The system of claim 2, wherein the buzz score is computed according to the equation:

$B=\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>\left({\mathrm{Rel}}_i·<br><br>F_i\right),$

where Rel is the computed numerical relevance score for a message, F is the computed numerical influence score for a poster of a message, and n is the number of messages in the set.

4. The system of claim 2, wherein the processing logic is adapted to compute a relative buzz score that measures relative changes in posting activity level, and the report at least indirectly reflects the relative buzz score.

5. The system of claim 1, wherein the processing logic further includes a textual analysis application that compares a content of the collected message with a plurality of known words and phrases indicative of expressions of sentiment to assign an opinion rating to the collected message, and the report at least indirectly reflects the opinion rating.

6. The system of claim 5, wherein the opinion rating is a numerical sentiment score on a numerical scale.

7. The system of claim 6, wherein the sentiment score is capable of being one of a positive sentiment score, a negative sentiment score and a zero sentiment score.

8. The system of claim 7, wherein the processing logic is further adapted to compute an aggregate positive sentiment score for a set of collected messages, and the report at least indirectly reflects the aggregate positive sentiment score.

9. The system of claim 8, wherein the aggregate positive sentiment score S_p is computed according to the equation:

$S_p=\underset{i=1}{\overset{m}{\sum <br><br>}}<br><br>s_{\mathrm{pm}},$

where m is the set of collected messages, and S_pm is a positive sentiment score for a message in the set of collected messages.

10. The system of claim 8, wherein the processing logic is further adapted to compute an aggregate negative sentiment score for a set of collected messages, and the report at least indirectly reflects the aggregate negative sentiment score.

11. The system of claim 10, wherein the aggregate negative sentiment score S_n is computed according to the equation:

$S_n=\underset{i=1}{\overset{m}{\sum <br><br>}}<br><br>s_{n<br><br><br><br>m},$

where m is the set of collected messages, and S_nm is a negative sentiment score for a message in the set of collected messages.

12. The system of claim 6, wherein the sentiment score is a net sentiment score.

13. The system of claim 12, wherein the processing logic is further adapted to compute an aggregate net sentiment score for a set of collected messages, and the report at least indirectly reflects the net sentiment score.

14. The system of claim 13, wherein the aggregate net sentiment score S is computed according to the equation:

$S=\underset{i=1}{\overset{m}{\sum <br><br>}}<br><br>\left(s_{\mathrm{pm}}-s_{n<br><br><br><br>m}\right),$

where m is the set of collected messages, S_pm is a positive sentiment score for a message in the set of collected messages, and S_nm is a negative sentiment score for the message.

15. A system as defined in claim 1, wherein the computer is a UNIX-based server.

16. A system for processing messages from a plurality of electronic discussion forums, comprising:

a computer implemented message collector for collecting messages from the plurality of electronic discussion forums;

a computer including processing logic to process the messages in order to track a plurality of pseudonyms based on a topic, wherein the processing logic computes a numerical relevance score that rates the relevance of a collected message posted by one pseudonym in the plurality of pseudonyms to the topic on a numerical scale, and wherein the processing logic computes a buzz score for a set of collected messages, wherein the buzz score is a measure of a level of activity within one or more of the electronic discussion forums and wherein the numerical relevance score is determined based on a set of rules, the rules having a plurality of conditions defining information relevant to the least one topic and having an associated weighting to indicate the strength a particular condition should have in determining the relevance of the messages with respect to the topic; and

a computer implemented report generator for generating a report on, at least in part, the topic using results from the processing logic so that the report at least indirectly reflects the buzz score.

17. The system of claim 16, wherein the processing logic is further adapted to compute a numerical migration score based on the set of messages and movement of influential posting activity within the one or more electronic discussion forums, and the report at least indirectly reflects the numerical migration score.

18. The system of claim 17, wherein the migration score provides a measurement, on a numerical scale, of the degree of movement of posting activity levels between topics or groups of topics from the topic.

19. The system of claim 17, wherein the migration score provides a measurement, on a numerical scale, of the degree of movement of posting activity levels between topics or groups of topics from the topic that is weighted by numerical influence scores of posting pseudonyms.

20. The system of claim 19, wherein the migration score includes an indication of direction of the movement.

21. The system of claim 19, wherein the migration score is computed based on a change in buzz levels.

22. The system of claim 21, wherein the migration score is computed based on the equation:

M_n=[B_n−

B<sub>n−

1</sub>], wherein B_n is the buzz level, on a numerical scale, for day n, and B<sub>n−

1</sub> is the buzz level, on a numerical scale, for the previous day.

23. A system as defined in claim 16, wherein the computer is a UNIX-based server.

24. A system for tracking message activity levels in a plurality of electronic discussion forums, comprising:

a database storing a series of topics, wherein each of the electronic discussion forums is associated with a topic from the series of topics;

a computer implemented message collector for collecting messages from the plurality of electronic discussion forums; and

a computer including a message categorizer for computing numerical relevance scores of the messages, wherein at least one numerical relevance score is computed based on the associated topic for a source forum of a collected message and wherein the numerical relevance scores are determined based on a set of rules, the rules having a plurality of conditions defining information relevant to the series of topics and having an associated weighting to indicate the strength a particular condition should have in determining the relevance of the messages with respect to the series of topics; and

a computer implemented report generator for generating a report on the associated topic reflecting, at least indirectly, the numerical relevance scores;

wherein the at least one numerical relevance score rates the relevance of the collected message to the topic on a numerical scale;

wherein at least one of the message collector or the message categorizer is adapted to extract a pseudonym author of the collected message to track a plurality of pseudonyms;

wherein at least one of the message collector or the message categorizer is adapted to compute a numerical influence score for the extracted pseudonym author, and wherein the influence score rates the influence of the extracted pseudonym author on a numerical scale, and the report at least indirectly reflects the numerical influence score; and

wherein at least one of the message collector and the message categorizer is adapted to compute a numerical buzz score for a topic or group of topics from the series of topics, and the report at least indirectly reflects the numerical buzz score.

25. The system of claim 24, wherein the database comprises a table that links forums to associated topics.

26. The system of claim 24, wherein the at least one numerical relevance score comprises multiple relevance scores corresponding to multiple associated topics, and wherein the highest one of the multiple relevance scores is maintained and the other of the multiple relevance scores are discarded or ignored.

27. The system of claim 24, wherein the extracted pseudonym author is analyzed to determine if the extracted pseudonym author is a new author or a previously-tracked author.

28. The system of claim 27, wherein the determination is made by comparing the extracted pseudonym author to an author'"'"'s table that stores a plurality of previously-tracked authors.

29. The system of claim 24, wherein the influence score is based at least in part on the impact of prior postings by the extracted pseudonym author on other pseudonym authors.

30. The system of claim 24, wherein at least one of the message collector or the message categorizer is adapted to compute a numerical reputation score for the extracted pseudonym author, and wherein the reputation score rates the extracted pseudonym author'"'"'s reputation on a numerical scale, and the report at least indirectly reflects the numerical reputation score.

31. The system of claim 30, wherein the reputation score is based on a plurality of influence scores for the extracted pseudonym author under multiple local pseudonyms that have been associated into a universal pseudonym.

32. The system of claim 24, wherein at least one of the message collector or the message categorizer is adapted to compute a relative numerical buzz score for the topic or the group of topics, and the report at least indirectly reflects the relative numerical buzz score.

33. The system of claim 32, wherein the relative buzz score is compared to an anomaly threshold in order to identify unusual discussion patterns.

34. The system of claim 24 or claim 32, wherein the electronic discussion forums include forums related to one or more of financial markets, stock exchanges, financial instruments and financial information.

35. The system of claim 24 or claim 32, wherein the series of topics includes one or more of financial markets and financial instruments.

36. The system of claim 24 or claim 32, wherein the group of topics corresponds to a stock index, and wherein the buzz score or the relative buzz score is computed by aggregating the buzz scores for the stocks making up the stock index.

37. The system of claim 24 or claim 32, wherein the group of topics corresponds to a market sector, and wherein the buzz score or the relative buzz score is computed by aggregating the buzz scores for stocks making up the sector.

38. The system of claim 24 or claim 32, wherein the series of topics includes a plurality of movies or a plurality of television shows.

39. The system of claim 24, wherein at least one of the message collector or the message categorizer is adapted to compute a numerical migration score for the topic or the group of topics, and wherein the migration score is computed on a numerical scale based on a change in buzz scores for the topic or group of topics, and the report at least indirectly reflects the numerical migration score.

40. A system as defined in claim 24, wherein the computer is a UNIX-based server.

41. A method for processing message traffic from a plurality of electronic discussion forums, comprising:

collecting messages from the plurality of electronic discussion forums;

processing with a processor the messages based on a topic in order to track a plurality of pseudonyms, wherein processing the messages comprises computing a numerical relevance score that rates the relevance of a collected message to the topic on a numerical scale, and wherein the numerical relevance score is determined based on a set of predefined rules, the predefined rules having a plurality of conditions defining information relevant to the topic and having an associated weighting to indicate the strength a particular condition should have in determining the relevance of the messages with respect to the topic; and

storing a report in a tangible memory regarding, at least in part, the topic, wherein a report output is based upon, at least in part, the numerical relevance score of a collected message and characteristics of the plurality of pseudonyms;

wherein processing the messages further comprises computing a numerical buzz score based on a set of messages for the topic, wherein the report output is also based upon, at least in part, the numerical buzz score.

42. The method of claim 41, wherein the buzz score measures posting activity level, on a numerical scale, weighted by the degree of relevance of the set of messages and the degree of influence of posting pseudonyms that posted the set of messages.

43. The method of claim 42, wherein processing the messages comprises computing a relative buzz score that measures relative changes in the posting activity level, wherein the report output is also based upon, at least in part, the relative buzz score.

44. The method of claim 41, wherein the buzz score is computed according to the equation:

$B=\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>\left({\mathrm{Rel}}_i·<br><br>F_i\right),$

where Rel is a computed numerical relevance score for a message, F is a computed numerical influence score for a poster of a message, and n is the number of messages in the set.

45. A method for processing messages from a plurality of electronic discussion forums, comprising:

collecting messages from the plurality of electronic discussion forums;

processing with a processor the messages based on a topic, wherein processing the messages comprises computing a numerical relevance score that rates the relevance of a collected message to the topic on a numerical scale, the numerical relevance score being determined based on a set of predefined rules, the predefined rules having a plurality of conditions defining information relevant to the topic and having an associated weighting to indicate the strength a particular condition should have in determining the relevance of the messages with respect to the topic, and wherein processing the messages comprises computing a buzz score for a set of collected messages, the buzz score being a measure of a level of activity within one or more of the electronic discussion forums; and

storing a report in a tangible memory on, at least in part, the topic, wherein a report output is based upon, at least in part, the numerical relevance score and the buzz score.

46. The method of claim 45, wherein processing the messages further comprises computing a numerical migration score based on the set of messages and movement of influential posting activity within the one or more electronic discussion forums, and the report output is further based upon, at least in part, the numerical migration score.

47. The method of claim 46, wherein the migration score provides a measurement, on a numerical scale, of the degree of movement of posting activity levels between topics or groups of topics from the topic.

48. The method of claim 47, wherein the migration score includes an indication of direction of the movement.

49. The method of claim 46, wherein the migration score provides a measurement, on a numerical scale, of the degree of movement of posting activity levels between topics or groups of topics from the topic that is weighted by influence scores of posting pseudonyms.

50. The method of claim 47, wherein the migration score is computed based on a change in the buzz score.

51. A method for processing messages from a plurality of electronic discussion forums, comprising:

collecting messages from the plurality of electronic discussion forums; and

processing with a processor the messages based on a series of topics in order to track a plurality of pseudonyms, wherein processing the messages comprises computing a relevance score for a collected message based on a topic and wherein the relevance score is determined based on a set of rules, the rules having a plurality of conditions defining information relevant to the topic and having an associated weighting to indicate the strength a particular condition should have in determining the relevance of the messages with respect to the topic, and wherein processing the messages comprises computing a buzz score for a set of collected messages;

wherein processing the messages further comprises computing a migration score based on the set of messages;

wherein the migration score provides a measurement of the movement of posting activity levels between topics or groups of topics from the series of topics;

wherein the migration score is computed based on a change in buzz levels;

wherein the migration score is computed based on the equation;

M_n=[B_n−

B<sub>n−

1</sub>], wherein B_n is the buzz level for day n and B<sub>n−

1</sub> is the buzz level for the previous day;

the method further comprising storing a report in a tangible memory on, at least in part, the topic, wherein a report output is based upon, at least in part, the migration score.

52. The method of claim 51, wherein processing the messages further comprises computing a sentiment score measuring a degree in which the collected message exhibits one or more of (a) positive sentiment, (b) negative sentiment and (c) no sentiment associated with the topic wherein the report output is further based upon, at least in part, the sentiment score.

53. A method for tracking message activity levels in a plurality of electronic discussion forums, comprising:

storing a series of topics in a database, wherein each of the electronic discussion forums is associated with a topic from the series of topics;

collecting messages from the plurality of electronic discussion forums;

computing with a processor numerical relevance scores of the messages, wherein at least one of the numerical relevance scores is computed based on the associated topic for a source forum of a collected message, and wherein each of the numerical relevance scores rates the relevance of the corresponding collected message to the topic on a numerical scale and wherein each of the numerical relevance scores is determined based on a set of rules, the rules having a plurality of conditions defining information relevant to the associated topic and having an associated weighting to indicate the strength a particular condition should have in determining the relevance of the messages with respect to the associated topic;

extracting a pseudonym author of at least one collected message to track a plurality of pseudonyms;

computing a numerical influence score for the extracted pseudonym author;

storing a report in a tangible memory regarding, at least in part, the topic, wherein a report output is based upon, at least in part, at least one of the numerical relevance scores and at least one of the numerical influence scores; and

computing a numerical buzz score for a topic or group of topics from the series of topics, wherein the numerical buzz score is a measure of a level of activity within one or more of the electronic discussion forums and the report output is further based upon, at least in part, the numerical buzz score.

54. The method of claim 53, further comprising linking forums to associated topics in a table in the database.

55. The method of claim 53, wherein the at least one relevance score comprises multiple relevance scores corresponding to multiple associated topics, and wherein the highest one of the multiple relevance scores is maintained and the other of the multiple relevance scores are discarded or ignored.

56. The method of claim 53, further comprising analyzing the extracted pseudonym author to determine if the extracted pseudonym author is a new author or a previously-tracked author.

57. The method of claim 56, wherein the determination is made by comparing the extracted pseudonym author to an author'"'"'s table that stores a plurality of previously-tracked authors.

58. The method of claim 53, wherein the influence score is based at least in part on an impact of prior postings by the extracted pseudonym author on other pseudonym authors.

59. The method of claim 53, further comprising computing a numerical reputation score for the extracted pseudonym author, wherein the report output is further based upon, at least in part, the numerical reputation score.

60. The method of claim 59, wherein the reputation score is based on a plurality of influence scores for the extracted pseudonym author under multiple local pseudonyms that have been associated into a universal pseudonym.

61. The method of claim 53, further comprising computing a relative buzz score for the topic or the group of topics, wherein the report output is further based upon, at least in part, the relative buzz score.

62. The method of claim 61, wherein the relative buzz score is compared to an anomaly threshold in order to identify unusual discussion patterns.

63. The method of claim 53 or claim 61, wherein the electronic discussion forums include forums related to one or more of financial markets, stock exchanges, financial instruments and financial information.

64. The method of claim 53 or claim 61, wherein the series of topics includes one or more of financial markets and financial instruments.

65. The method of claim 53 or claim 61, wherein the group of topics corresponds to a stock index, and wherein the buzz score or the relative buzz score is computed by aggregating the buzz scores for the stocks making up the stock index.

66. The method of claim 53 or claim 61, wherein the group of topics corresponds to a market sector, and wherein the buzz score or the relative buzz score is computed by aggregating the buzz scores for stocks making up the sector.

67. The method of claim 53 or claim 61, wherein the series of topics includes a plurality of movies or a plurality of television shows.

68. The method of claim 53 further comprising computing a numerical migration score for the topic or the group of topics, wherein the migration score is computed based on a change in buzz scores for the topic or group of topics, and wherein the report output is further based upon, at least in part, the migration score.

69. A computer implemented method for tracking and reporting electronic message activity in one or more electronic discussions, comprising:

collecting by a computer system a plurality of messages from one or more electronic discussions;

for each message, processing the message to track a plurality of pseudonyms by computing by the computer system a numerical relevance score for the message posted by one of the plurality of pseudonyms based upon, at least in part, the relevance of the message to a topic, wherein the numerical relevance score of each message is computed on a numerical scale, and wherein the numerical relevance score of each message is determined based on a set of predefined rules, the predefined rules having a plurality of conditions defining information relevant to the topic and having an associated weighting to indicate the strength a particular condition should have in determining the relevance of the messages with respect to the topic;

for each message, determining by the computer system the presence of at least one of a positive sentiment, a negative sentiment and an absence of sentiment;

compiling by the computer system analysis results based upon, at least in part, the topic, the relevance scores for the plurality of messages and the determination of the presence of at least one of the positive sentiment, negative sentiment and absence of sentiment in each message;

generating by the computer system an electronic message activity report using the analysis results;

computing a numerical buzz score for the topic, wherein the report at least indirectly reflects the numerical buzz score.

70. The computer implemented method of claim 69, wherein determining the presence of at least one of the positive sentiment, negative sentiment and absence of sentiment in each message includes analyzing by the computer system textual patterns in the message for the expression of sentiment.

71. The computer implemented method of claim 70, wherein analyzing textual patterns in the message for the expression of sentiment includes consulting a sentiment model that includes words used for expressing positive and negative sentiment for the topic.

72. The computer implemented method of claim 71, wherein the sentiment model is a financial sentiment model.

73. The computer implemented method of claim 71, wherein the sentiment model is an entertainment sentiment model.

74. The computer implemented method of claim 69, wherein determining the presence of at least one of the positive sentiment, negative sentiment and absence of sentiment in each message includes computing a numerical sentiment rating on a numerical scale and compiling the analysis results based further upon the numerical sentiment ratings for the plurality of messages.

75. The computer implemented method of claim 74, wherein computing a numerical sentiment rating on a numerical scale computes both a positive sentiment rating and a negative sentiment rating, and compiling the analysis results is based farther upon the positive and negative sentiment ratings for the plurality of messages.

76. The computer implemented method of claim 74, further comprising aggregating by the computer system the sentiment ratings for the plurality of messages, and compiling the analysis results is based further upon the aggregated sentiment ratings.