System And Method For Identifying Topics For Short Text Communications

• US 20110153595A1
• Filed: 12/23/2009
• Published: 06/23/2011
• Est. Priority Date: 12/23/2009
• Status: Active Grant

• Current Assignee: Palo Alto Research Center, Inc. (Xerox Holdings Corp.)
• Original Assignee: Palo Alto Research Center, Inc. (Xerox Holdings Corp.)
• Inventors: Bernstein, Michael, Heng, Lichan, Suh, Bengwon, Chi, Ed H.
• Granted Patent: US 8,725,717 B2
• Time in Patent Office: Days
• US Class Current: 707/722
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06Q 10/107 Computer-aided management of electronic mailing [e-mailing]

Abstract

A system and method for identifying topics for short text communications is provided. Tokens are extracted from a short text communication. A query is generated using the extracted tokens. The query is applied to a set of documents. Those documents in the set that match the query are identified as search results. Salient terms associated with each of the search results are identified. A threshold is applied to the identified salient terms. The salient terms that satisfy the threshold are selected as topics for the short text communication.

Claims

1. A system for identifying topics for short text communications, comprising:

an extraction module to extract tokens from a short text communication;

a query module to generate a query using the extracted tokens and to apply the query to a set of documents;

an result identification module to identify those documents in the set that match the query as search results;

a threshold module to identify salient terms associated with each of the search results and to apply a threshold to the identified salient terms; and

a topic module to select the salient terms that satisfy the threshold as topics for the short text communication.

2. A system according to claim 1, further comprising:

a part-of-speech tagger to identify the tokens in the short text communication.

3. A system according to claim 1, further comprising:

a concatenation module to concatenate the extracted nouns to form the query.

4. A system according to claim 1, wherein the topics comprise at least one of a token identified in the short text communication and a token absent from the short text communication.

5. A system according to claim 1, wherein the salient terms are identified at least one of automatically or by frequency criteria.

6. A system according to claim 1, wherein the set of documents comprise Web-based documents.

7. A system according to claim 1, further comprising:

a display to present the topics with other topics that are associated with a set of short text communications; and

a filter module to filter the set of short text communications based on one or more of the presented topics.

8. A system according to claim 1, further comprising:

a topic identification module to identify the topics, comprising;

a vote module to assign votes to each salient term when that salient term appears in one of the search results;

a vote threshold module to total the votes for each salient term and to apply a threshold to the total votes; and

a topic designation module to designate the salient terms that satisfy the threshold as the topics.

9. A system according to claim 8, further comprising:

a comparison module to apply a lower threshold to the salient terms that appear within the short text communication.

10. A system according to claim 8, further comprising:

a suggestion module to retain the salient terms that fail to satisfy the threshold as suggestions for the short text communication.

11. A method for identifying topics for short text communications, comprising:

extracting tokens from a short text communication;

generating a query using the extracted tokens and applying the query to a set of documents;

identifying those documents in the set that match the query as search results;

identifying salient terms associated with each of the search results and applying a threshold to the identified salient terms; and

selecting the salient terms that satisfy the threshold as topics for the short text communication.

12. A method according to claim 11, further comprising:

applying a part-of-speech tagger to the short text communication to identify the tokens.

13. A method according to claim 11, further comprising:

concatenating the extracted nouns to form the query.

14. A method according to claim 11, wherein the topics comprise at least one of a token identified in the short text communication and a token absent from the short text communication.

15. A method according to claim 11, wherein the salient terms are identified at least one of automatically or by frequency criteria.

16. A method according to claim 11, wherein the set of documents comprise Web-based documents.

17. A method according to claim 11, further comprising:

presenting the topics in a display with other topics; and

filtering a set of short text communications based on one or more of the presented topics.

18. A method according to claim 11, further comprising:

identifying the topics, comprising;

assigning votes to each salient term when that salient term appears in one of the search results;

totaling the votes for each salient term and applying a threshold to the total votes; and

designating the salient terms that satisfy the threshold as the topics.

19. A method according to claim 18, further comprising:

applying a lower threshold to the salient terms that appear within the short text communication.

20. A method according to claim 18, further comprising:

retaining the salient terms that fail to satisfy the threshold as suggestions for the short text communication.