SYSTEM AND METHOD FOR ALIGNING MESSAGES TO AN EVENT BASED ON SEMANTIC SIMILARITY

• US 20130081056A1
• Filed: 02/10/2012
• Published: 03/28/2013
• Est. Priority Date: 09/23/2011
• Status: Active Grant

• Current Assignee: Avaya Incorporated
• Original Assignee: Avaya Incorporated
• Inventors: JOHN, Ajita, SELIGMANN, Doree Duncan, HU, Yuheng
• Granted Patent: US 11,055,334 B2
• Time in Patent Office: Days
• US Class Current: 719/313
• CPC Class Codes: G06F 16/35 Clustering; Classification

Abstract

Disclosed are systems, methods, and non-transitory computer-readable storage media for analyzing an event via social media. A system configured to practice the method segments a document associated with an event based on a plurality of inferred topics to yield a plurality of segments. Next, the system assigns each message in a corpus of messages associated with the event to one of the inferred topics to yield assigned topics. The messages can include at least one of a status update, a tweet, a blog post, a tag, a chat message, a comment, and a check-in. Moreover, the messages can originate from different social networks. Based on the assigned topics, the system then associates each message in the corpus of messages with at least one of the document, a segment from the plurality of segments, and a segment of the event.

Claims

1. A method comprising:

segmenting, via a processor, a document associated with an event based on a plurality of inferred topics to yield a plurality of segments;

assigning each message in a corpus of messages associated with the event to one of the inferred topics to yield assigned topics; and

based on the assigned topics, associating each message in the corpus of messages with at least one of the document, a segment from the plurality of segments, and a segment of the event.

2. The method of claim 1, wherein associating each message in the corpus of messages with at least one of the document, the segment from the plurality of segments, and the segment of the event is also based on a joint statistical model of the corpus of messages and at least one of the document and the event.

3. The method of claim 1, wherein segmenting the document is based on a Latent Dirichlet Allocation analysis.

4. The method of claim 1, wherein assigning each message in the corpus of messages to one of the inferred topics is based on a Latent Dirichlet Allocation analysis.

5. The method of claim 1, further comprising:

ranking the plurality of segments based on how many messages in the corpus of messages are associated with each of the plurality of segments.

6. The method of claim 1, wherein each message in the corpus of messages comprises at least one of a status update, a tweet, a blog post, a tag, a chat message, a comment, and a check-in.

7. The method of claim 1, wherein associating each message in the corpus of messages to at least one of the document, the segment from the plurality of segments, and the segment of the event further comprises:

calculating a term-vector similarity for each message in the corpus of messages to each of the plurality of segments.

8. The method of claim 1, further comprising filtering out noisy terms above a noise threshold from a respective message in the corpus of messages.

9. The method of claim 1, further comprising associating words in each message in the corpus of messages with at least one of a specific topic in the event and a general topic in the event.

10. The method of claim 1, wherein a message from the corpus of messages is associated with more than one of the plurality of segments.

11. The method of claim 1, wherein the corpus of messages comprises a combination of data from a plurality of social networks.

12. The method of claim 1, wherein the plurality of segments comprises at least one segment from a second document associated with the event.

13. The method of claim 1, wherein associating each message in the corpus of messages to at least one of the document, the segment from the plurality of segments, and the segment of the event further comprises inferring topical associations between the corpus of messages and at least one of the document and the event.

14. The method of claim 1, further comprising:

adding external information to a message from the corpus of messages; and

using at least part of the external information to assign the message to one of the inferred topics in the plurality of inferred topics.

15. The method of claim 1, further comprising:

identifying a topic of a message;

identifying additional information related to the topic; and

expanding the message based on the additional information.

17. The system of claim 15, wherein the corpus of messages comprises a combination of data from a plurality of social networks.

18. The system of claim 15, the computer-readable medium storing further instructions which, when executed by the processor, cause the processor to perform a method further comprising:

receiving a plurality of messages;

filtering the plurality of messages to yield a set of messages that are descriptive of the event; and

identifying episodal topics in the event based on the set of messages.

20. The non-transitory computer-readable storage medium of claim 18, wherein the augmented version of the event comprises at least one of a time-based display and a topic-based display.

16. A system comprising:

a processor; and

a computer-readable medium having stored therein instructions which, when executed by the processor, cause the processor to perform a method comprising;

segmenting a document associated with an event based on a plurality of inferred topics to yield a plurality of segments, wherein a number of inferred topics in the plurality of inferred topics is based on a tuning threshold;

assigning each message in a corpus of messages associated with the event to one of the inferred topics to yield assigned topics; and

based on the assigned topics, associating each message in the corpus of messages with at least one of the document, a segment from the plurality of segments, and a segment of the event.

19. A non-transitory computer-readable storage medium storing instructions which, when executed by a computing device, cause the computing device to perform steps comprising:

segmenting a document associated with an event based on a plurality of inferred topics to yield a plurality of segments;

assigning each message in a corpus of messages associated with the event to one of the inferred topics to yield assigned topics;

based on the assigned topics, associating each message in the corpus of messages with at least one of the document, a segment from the plurality of segments, and a segment of the event; and

playing an augmented version of the event, wherein the augmented version of the event comprises at least part of the event and a topically related message from the corpus of messages.