METHODS, APPARATUS, AND COMPUTER-READABLE MEDIA FOR ANALYZING CONVERSATIONAL-TYPE DATA

• US 20080306899A1
• Filed: 06/07/2007
• Published: 12/11/2008
• Est. Priority Date: 06/07/2007
• Status: Abandoned Application

• Current Assignee: Battelle Memorial Institute
• Original Assignee: Battelle Memorial Institute
• Inventors: Rose, Stuart J., Schur, Anne, Gregory, Michelle L., Love, Douglas V.
• Application Number:
  US11/759,803
• Publication Number:
  US 20080306899A1
• Time in Patent Office: Days
• US Class Current: 707/1
• CPC Class Codes: G06F 16/345 Summarisation for human users

Abstract

Methods, apparatus, and computer-readable media for analyzing conversational-type data by association of two or more types of extracted information in view of time are disclosed according to some aspects. In one embodiment, analysis of conversational-type data comprises identification of topical segments within the conversational-type data and linking of the topical segments with at least one other type of pertinent, extracted information. The linking can be based on a sequential order of utterances that compose, at least in part, the conversational-type data.

Claims

1. A computer-implemented method for analysis of conversational-type data by association of two or more types of extracted information in view of time, the method comprising

automatically identifying topical segments within the conversational-type data, wherein the conversational-type data comprises a plurality of utterances occurring in a time period;

and

linking the topical segments with at least one other type of extracted information from the conversational-type data, wherein the linking is based, at least in part, on a sequential order of the utterances.

2. The computer-implemented method as recited in claim 1, wherein the conversational-type data comprises static data, streaming data, data streaming in near-real time, or combinations thereof.

3. The computer-implemented method as recited in claim 1, occurring in near real-time, for conversational-type data comprising streaming data.

4. The computer-implemented method as recited in claim 1, wherein conversational-type data comprises utterances generated by a plurality of participants engaged in a dialogue.

5. The computer-implemented method as recited in claim 4, wherein the conversational-type data is selected from the group consisting of chat logs, phone transcripts, meeting transcripts, instant messaging, usenet groups, and combinations thereof.

6. The computer-implemented method as recited in claim 1, wherein the conversational-type data is a blog or email correspondence.

7. The computer-implemented method as recited in claim 4, wherein the conversational-type data further comprises a sequence position and a participant name for each utterance, the utterances being arranged according to the sequence position.

8. The computer-implemented method as recited in claim 7, wherein automatically identifying topical segments comprises determining cohesion among the elements in the conversational-type data, the cohesion being based, at least in part, on associations among the utterances over a range of sequence positions.

9. The computer-implemented method as recited in claim 7, wherein automatically identifying topical segments comprises applying a windowless technique for topic segmentation.

10. The computer-implemented method as recited in claim 8, wherein determining cohesion comprises

quantifying the similarity between each neighboring pair of utterances;

and

iteratively joining the most similar neighboring pair, recording cohesion of the most similar neighboring pair, and re-quantifying similarities of neighboring elements to the most similar neighboring pair.

11. The computer-implemented method as recited in claim 10, wherein said quantifying is based on utterance vectors of the elements, each utterance vector being a function or aggregation of term vectors describing the similarity of a given term with selected features in the conversational-type data.

12. The computer-implemented method as recited in claim 11, wherein each term vector comprises correlations between one term and each of the remaining terms, and determination of the correlations comprises:

identifying all positions for terms in a pair of terms;

generating an array of all unique positions of the terms in the pair;

generating a paired value array for each term in the pair of terms, wherein for each unique position of one of the paired terms, the next closest position of either term is recorded in its respective paired value array; and

providing the paired value arrays to a correlation function.

13. The computer-implemented method as recited in claim 1, wherein the other type of extracted information comprises named entities.

14. The computer-implemented method as recited in claim 1, wherein the other type of extracted information comprises participants involved in generation of the conversational-type data.

15. The computer-implemented method as recited in claim 14, wherein the linking comprises mapping the participants to the topical segments over a period of time.

16. The computer-implemented method as recited in claim 1, wherein the other type of extracted information comprises participant attitude.

17. The computer-implemented method as recited in claim 16, further comprising linking participants involved in generation of the conversational-type data, the participant attitude, and the topical segments, one with another.

18. The computer-implemented method as recited in claim 1, wherein the other type of extracted information comprises participant roles.

19. The computer-implemented method as recited in claim 1, wherein the topical segments and two or more other types of extracted information are linked and the other types of extracted information are selected from the group consisting of participant attitude, named entities, participants, and participant roles.

20. The computer-implemented method as recited in claim 1, wherein the linking based on a sequential order comprises determining links between the topical segments and the other types of extracted information for a given portion of the sequential order.

21. The computer-implemented method as recited in claim 1, further comprising representing the linking between the topical segments and the other types of extracted information on a display device.

22. The computer-implemented method as recited in claim 21, wherein the representing comprises generating a matrix-based representation.

23. The computer-implemented method as recited in claim 22, wherein at least one dimension of the matrix-based representation comprises a representation of the sequential order.

24. The computer-implemented method as recited in claim 22, further comprising updating the matrix-based representation in real time, or near-real time.

25. The computer-implemented method as recited in claim 1, further comprising generating a message, alert signal, or combination thereof when aspects of the linking between the topical segments and the other types of extracted information satisfy one or more predetermined criteria.

26. An apparatus for analysis of conversational-type data comprising a plurality of utterances, the apparatus comprising processing circuitry configured to identify and link topical segments and at least one other type of extracted information within the conversational-type data, wherein the linking is based at least in part, on a sequential order of the utterances.

27. The apparatus as recited in claim 26, wherein processes executed by the processing circuitry are arranged according to a modular architecture comprising a central processing engine and a plurality of processing components, wherein processing components are called by the central processing engine and the central processing engine provides input to, and collects output from, each component.

28. The apparatus as recited in claim 27, wherein the processing components comprise software modules causing the processing circuitry to perform processes selected from the group consisting of topic segmentation, sentiment analysis, named entity extraction, and participant information analysis.

29. The apparatus as recited in claim 26, further comprising a user interface operably connected to the processing circuitry and configured to display a representation of the linking between the topical segments and the other types of extracted information.

30. The apparatus as recited in claim 29, wherein the representation comprises a matrix-based representation.

31. The apparatus as recited in claim 29, wherein at least one dimension of the matrix-based representation comprises a representation of the sequential order.

32. A computer-readable medium having stored thereon a data structure comprising:

one or more fields containing data representing topical segments within conversational-type data, wherein the conversational-type data comprises a plurality of utterances;

one or more fields containing data representing other types of extracted information from the conversational-type data; and

one or more fields containing data representing a portion of a sequential order of the utterances over which the topical segments and the other types of extracted information are defined, wherein the topical segments and the other types of extracted information are linked, one with another, based, at least in part, on the sequential order of the utterances.