ANALYZING PARALLEL TOPICS FROM CORRELATED DOCUMENTS

• US 20110202484A1
• Filed: 02/18/2010
• Published: 08/18/2011
• Est. Priority Date: 02/18/2010
• Status: Abandoned Application

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Sun, Jimeng, Bose, Abhijit, Zhang, Duo, Anerousis, Nikolaos
• Application Number:
  US12/708,053
• Publication Number:
  US 20110202484A1
• Time in Patent Office: Days
• US Class Current: 706/12
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks

Abstract

Access is obtained to a parallel corpus including a problem corpus and a solution corpus. A first plurality of topics are mined from the problem corpus and a second plurality of topics are mined from the solution corpus. A transition probability from the first plurality of topics to the second plurality of topics is determined, to identify a most appropriate one of the topics from the solution corpus for a given one of the topics from the problem corpus.

Claims

1. A method comprising:

obtaining access to a parallel corpus comprising a problem corpus and a solution corpus;

mining a first plurality of topics from said problem corpus;

mining a second plurality of topics from said solution corpus; and

determining transition probability from said first plurality of topics to said second plurality of topics to identify a most appropriate one of said topics from said solution corpus for a given one of said topics from said problem corpus.

2. The method of claim 1, wherein said determining step employs expectation maximization.

3. The method of claim 1, wherein, in said mining steps, there is a one-to-one correspondence between said first plurality of topics and said second plurality of topics.

4. The method of claim 1, further comprising incorporating prior knowledge regarding at least selected ones of said first plurality of topics, in said determining step, using a maximum a posteriori technique.

5. The method of claim 4, wherein said incorporating comprises overweighting important words.

6. The method of claim 4, wherein, in said incorporating step, said selected ones of said first plurality of topics have manually assigned categories.

7. The method of claim 1, further comprising generating said parallel corpus by accumulating problem tickets during provision of information technology support for a computer system.

8. The method of claim 7, further comprising:

selecting a subset of said tickets as most representative of a given one of said first plurality of topics; and

displaying said selected subset to a human expert.

9. The method of claim 1, further comprising providing a system, wherein the system comprises distinct software modules, each of the distinct software modules being embodied on a computer-readable storage medium, and wherein the distinct software modules comprise a classification module and a diagnosis module;

wherein;

said mining of said first plurality of topics is carried out by said classification module executing on at least one hardware processor;

said mining of said second plurality of topics is carried out by said diagnosis module executing on said at least one hardware processor; and

said determining step is carried out by said diagnosis module executing on said at least one hardware processor.

10. A computer program product comprising a computer readable storage medium having computer readable program code embodied therewith, the computer readable program code comprising:

computer readable program code configured to obtain access to a parallel corpus comprising a problem corpus and a solution corpus;

computer readable program code configured to mine a first plurality of topics from said problem corpus;

computer readable program code configured to mine a second plurality of topics from said solution corpus; and

computer readable program code configured to determine transition probability from said first plurality of topics to said second plurality of topics to identify a most appropriate one of said topics from said solution corpus for a given one of said topics from said problem corpus.

11. The computer program product of claim 10, wherein said computer readable program code configured to determine employs expectation maximization.

12. The computer program product of claim 10, wherein there is a one-to-one correspondence between said first plurality of topics and said second plurality of topics.

13. The computer program product of claim 10, further comprising computer readable program code configured to incorporate prior knowledge regarding at least selected ones of said first plurality of topics, in said computer readable program code configured to determine, using a maximum a posteriori technique.

14. The computer program product of claim 13, wherein said computer readable program code configured to incorporate overweights important words.

15. The computer program product of claim 13, wherein said selected ones of said first plurality of topics have manually assigned categories.

16. The computer program product of claim 10, further comprising computer readable program code configured to generate said parallel corpus by accumulating problem tickets during provision of information technology support for a computer system.

17. The computer program product of claim 16, further comprising:

computer readable program code configured to select a subset of said tickets as most representative of a given one of said first plurality of topics; and

computer readable program code configured to display said selected subset to a human expert.

18. An apparatus comprising:

a memory; and

at least one processor, coupled to said memory, and operative to;

obtain access to a parallel corpus comprising a problem corpus and a solution corpus;

mine a first plurality of topics from said problem corpus;

mine a second plurality of topics from said solution corpus; and

determine transition probability from said first plurality of topics to said second plurality of topics to identify a most appropriate one of said topics from said solution corpus for a given one of said topics from said problem corpus.

19. The apparatus of claim 18, wherein said at least one processor is operative to determine by employing expectation maximization.

20. The apparatus of claim 18, wherein there is a one-to-one correspondence between said first plurality of topics and said second plurality of topics.

21. The apparatus of claim 18, wherein said at least one processor is further operative to incorporate prior knowledge regarding at least selected ones of said first plurality of topics, in said determining, using a maximum a posteriori technique.

22. The apparatus of claim 21, wherein said at least one processor is operative to incorporate by overweighting important words.

23. The apparatus of claim 21, wherein said selected ones of said first plurality of topics have manually assigned categories.

24. The apparatus of claim 18, further comprising a plurality of distinct software modules, each of the distinct software modules being embodied on a computer-readable storage medium, and wherein the distinct software modules comprise a classification module and a diagnosis module;

wherein;

said at least one processor is operative to mine said first plurality of topics by executing said classification module;

said at least one processor is operative to mine said second plurality by executing said diagnosis module; and

said at least one processor is operative to determine by executing said diagnosis module.

25. An apparatus comprising:

means for obtaining access to a parallel corpus comprising a problem corpus and a solution corpus;

means for mining a first plurality of topics from said problem corpus;

means for mining a second plurality of topics from said solution corpus; and

means for determining transition probability from said first plurality of topics to said second plurality of topics to identify a most appropriate one of said topics from said solution corpus for a given one of said topics from said problem corpus.