DYNAMIC CREATION OF TOPICAL KEYWORD TAXONOMIES

• US 20110276581A1
• Filed: 05/10/2010
• Published: 11/10/2011
• Est. Priority Date: 05/10/2010
• Status: Active Grant

• Current Assignee: Oracle OTC Subsidiary LLC (Oracle Corporation)
• Original Assignee: Oracle OTC Subsidiary LLC (Oracle Corporation)
• Inventors: Zelevinsky, Vladimir
• Granted Patent: US 9,208,435 B2
• Time in Patent Office: Days
• US Class Current: 707/766
• CPC Class Codes: G06F 16/3323 using document space presentation or visualization, e.g. category, hierarchy or range presentation and selection
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation

Abstract

System and methods for dynamically generating taxonomies of keywords and/or descriptors are provided. In one example, a navigation system for accessing a corpus of information provides for dynamic taxonomy generation expanding upon a topic entered by a user in a user interface. The navigation system generates, dynamically, at least one term associated with the received topic from at least one sense or meaning retrieved from a semantic network. The navigation system is further configured to present to the user the at least one term as a selectable refinement in response to receiving the topic entered by the user in the user interface. The system can also be configured to retrieve terms and/or senses from the semantic network and evaluate any retrieved terms for their informativeness. The system can further cache any information generated during taxonomy creation and update the corpus to reflect useful refinements.

Claims

1. A computer implemented method for generating refinements within a corpus, the method comprising:

receiving a topic from a user in a user interface;

generating, dynamically, at least one term associated with the topic, wherein the act of generating the at least one term includes acts of;

communicating a query to a semantic network based on the topic,

receiving the at least one term from the semantic network in response to the query; and

presenting automatically to the user in the user interface the at least one term as a selectable refinement in response to receiving the topic from the user entered in the user interface.

2. The method according to claim 1, wherein the act of receiving the at least one term from the semantic network includes an act of receiving at least one sense of the received topic, and wherein the at least one term corresponds to the at least one sense.

3. The method according to claim 1, further comprising an act of comparing the at least one term against a threshold value that identifies informativeness of the at least one term.

4. The method according to claim 3, further comprising an act of dynamically determining the threshold value.

5. The method according to claim 2, wherein the threshold value that identifies informativeness of the at least one term includes a value for at least one of precision, recall, f-measure, and a salience metric.

6. The method according to claim 2, wherein the act of receiving the at least one term from the semantic network includes an act of receiving at least one candidate term that corresponds to each sense of the at least one sense of the received topic.

7. The method according to claim 6, further comprising an act of identifying the at least one term from the at least one candidate term, wherein the act of identifying the at least one term includes acts of:

evaluating the at least one candidate term against a threshold value that identifies informativeness of the at least one term; and

determining that the at least one candidate term meets the threshold value.

8. The method according to claim 1, wherein the act of receiving the at least one term from the semantic network includes receiving a plurality of terms from the semantic network, wherein the plurality of terms correspond to a respective sense of a plurality of senses for the topic.

9. The method according to claim 8, further comprising an act of grouping the plurality of terms by the respective sense for each term.

10. The method according to claim 1, wherein the act of identifying at least one sense of the topic from the semantic network includes an act of restricting the identified at least one sense by semantic relationship.

11. The method according to claim 9, wherein the semantic relationship includes at least one of a meronym, holonym, and hyponym.

12. The method according to claim 1, further comprising an act of caching the generated at least one term.

13. The method according to claim 1, further comprising an act of updating the corpus with the at least one term.

14. The method according to claim 1, further comprising an acts of:

tracking a usage pattern associated with received topics; and

updating the corpus with the at least one term in response to the usage pattern indicating an established user demand.

15. The method according to claim 1, further comprising an act of associating a navigation operation with the at least one term.

16. A system for dynamically generating taxonomies, the system comprising:

a communication component configured to receive a topic from a user input in a user interface on a host computer system;

a generation component configured to dynamically generate at least one term associated with the received topic, wherein the generation component is further configured to;

communicate a query to a semantic network based on the received topic, and

receive the at least one term from the semantic network in response to the query; and

wherein the communication component is further configured to communicate to the host computer system the at least one term as a selectable refinement displayed in the user interface in response to receiving the topic from the user in the user interface.

17. The system according to claim 16, wherein the received at least one term corresponds to at least one sense of the received topic.

18. The system according to claim 16, further comprising an evaluation component configured to compare the at least one term against a threshold value that identifies informativeness of the at least one term.

19. The system according to claim 18, wherein the evaluation component is further configured to dynamically generate the threshold value.

20. The system according to claim 18, wherein the evaluation component calculates for the at least one term a value for at least one of precision, recall, f-measure, and a salience metric.

21. The system according to claim 16, wherein the generation component is further configured to retrieve at least one candidate term that corresponds to each sense of the at least one sense of the topic.

22. The system according to claim 21, wherein the evaluation component is further configured to:

compare the at least one candidate term against a threshold value that identifies informativeness of the at least one term; and

present the at least one candidate term as the at least one term in response to a determination that the at least one candidate term meets the threshold value.

23. The system according to claim 16, wherein the generation component is further configured to receive a plurality of terms from the semantic network, wherein the plurality of terms correspond to a respective sense of a plurality of senses for the topic.

24. The system according to claim 16, wherein generation component is configured to restrict the at least one term received by semantic relationship.

25. The system according to claim 16, further comprising a storage component configured to cache the generated at least one term.

26. The system according to claim 16, further comprising an update component configured to update the corpus with the at least one term.

27. The system according to claim 16, further comprising a storage component configured to:

track usage patterns associated with submitted topics; and

update the corpus with the at least one term in response to the usage pattern indicating a predetermined user demand.

28. The system according to claim 16, further comprising an evaluation component configured to automatically filter the at least one term retrieved.

29. The system according to claim 16, wherein the generation component is further configured to receive a plurality of terms including the at least one term from the semantic network, and the system further comprises an evaluation component configured to filter, automatically, at least one of the plurality of terms.