Granted Patent B2

US 8,655,882 · App. 13/221,974 · Granted Feb 18, 2014

Inventors: Donald R. Kretz (Plano, TX); William D. Phillips (Tampa, FL); Bruce E. Peoples (State College, PA); Justin W. Toennies (Largo, FL)

Assignee: Raytheon Company

Abstract

Claims (94)

1. A system for ontology candidate selection and comparison comprising:

a microprocessor;

an ontology candidate selection component executing on the microprocessor and configured to:

compare at least a portion of a plurality of ontology candidates based upon a candidate selection rule; and

based upon said comparison, select a pair of ontologies from the plurality of ontology candidates; and

an ontology similarity component coupled to the ontology candidate selection component and configured to:

generate a similarity outcome related to the pair of ontologies based upon a similarity rule; and

evaluate at least one of: the candidate selection rule or the similarity rule based upon the similarity outcome;

wherein the pair of ontologies includes a first ontology and a second ontology and the similarity rule is based upon:

a plurality of text-based confidence levels associated with a plurality of text-based associations, each text-based association including concept strings from the first ontology and the second ontology;

a text-based threshold confidence level;

a plurality of semantic confidence levels associated with a plurality of semantic-based associations, each semantic-based association including concept strings from the first ontology and the second ontology; and

a semantic-based threshold confidence level,

and the similarity outcome is based upon:

at least one semantic-based association associated with a semantic-based confidence level greater than or equal to the semantic-based threshold confidence level; and

at least one text-based association associated with a text-based confidence level greater than or equal to the text-based threshold confidence level.

2. The system of claim 1 , wherein the ontology candidate selection component is further configured to:

generate a set of ontology combinations from the plurality of ontologies, each ontology combination including a first ontology candidate and second ontology candidate;

compare the first ontology candidate and the second ontology candidate in each ontology combination based on the candidate selection rule;

generate a plurality of sorted comparisons of the ontology combinations; and

select the pair of ontologies according to the sorted comparisons.

3. The system of claim 1 , wherein the candidate selection rule is associated with at least one of: ontology observational criteria, ontology structural criteria, ontology functional criteria, ontology processing time criteria, or ontology usability criteria.

4. The system of claim 3 , wherein the ontology candidate selection component is further configured to:

select the pair of ontologies based on a comparison of at least one of: an ontology observational criterion of the first ontology and an ontology observational criterion of the second ontology or an ontology functional criterion of the first ontology and an ontology functional criterion of the second ontology.

5. The system of claim 1 , wherein the pair of ontologies includes a first ontology and a second ontology and the similarity rule is based on a confidence level associated with a text-based association including a first concept string from the first ontology and a second concept string from the second ontology.

6. The system of claim 5 , wherein the similarity rule is further based on a threshold confidence level and the similarity outcome is based on at least one text-based association associated with a confidence level greater than or equal to the threshold confidence level.

7. The system of claim 6 , wherein the ontology similarity component is further configured to:

generate the threshold confidence level based on an F-score associated with the at least one text-based association.

8. The system of claim 1 , wherein the pair of ontologies includes a first ontology and a second ontology and the similarity rule is based on a confidence level associated with a semantic-based association including a first concept from the first ontology and a second concept string from the second ontology.

9. The system of claim 8 , wherein the similarity rule is further based on a threshold confidence level and the similarity outcome includes at least one semantic-based association associated with a confidence level greater than or equal to the threshold confidence level.

10. The system of claim 9 , wherein the ontology similarity component is further configured to:

generate the threshold confidence level based on an F-score and a significant difference calculation associated with the at least one semantic-based association.

11. The system of claim 1 , wherein the similarity outcome is further based on a similarity outcome score of a number of text-based associations greater than or equal to the text-based threshold confidence level and a number of semantic-based associations greater than or equal to the semantic-based threshold confidence level.

12. The system of claim 11 , wherein the ontology similarity component is further configured to:

evaluate at least one of: the candidate selection rule or the similarity rule, said evaluation based on a plurality of sorted similarity outcome scores associated with a plurality of pairings of ontologies.

13. The system of claim 1 , wherein the ontology similarity component is further configured to:

evaluate at least one of: the candidate selection rule or the similarity rule, said evaluation based on a plurality of sorted similarity outcomes associated with a plurality of pairings of ontologies.

14. The system of claim 1 , further comprising:

an ontology alignment component coupled to the ontology similarity component and configured to:

generate an alignment mapping between similar concepts in the pair of ontologies based on the similarity outcome.

15. The system of claim 14 , wherein the ontology alignment component further generates a new or modified ontology, schema, or relationship list based on the alignment mapping.

16. A system for ontology candidate selection and comparison comprising:

a microprocessor;

a candidate evaluation agent executing on the microprocessor and configured to:

receive ontology data associated with a plurality of ontology candidates;

compare at least one combination of the ontology candidates based on a candidate selection rule to generate a comparison outcome; and

select a pair of ontologies based on the comparison outcome;

a similarity agent configured to:

generate a similarity outcome based on a similarity rule applied to the selected pair of ontologies; and

a heuristic agent coupled to the candidate evaluation agent and the similarity agent and configured to:

select the candidate selection rule from a plurality of candidate selection rules;

select the similarity rule from a plurality of ontology similarity rules; and

evaluate at least one of: the candidates selection rule or the similarity rule based on the similarity outcome,

wherein the similarity rule is further based on a text-based threshold confidence level and a semantic-based threshold confidence level and the similarity outcome is based on:

at least one semantic-based association associated with a semantic-based confidence level greater than or equal to the semantic-based threshold confidence level; and

at least one text-based association associated with a text-based confidence level greater than or equal to the text-based threshold confidence level.

17. The system of claim 16 , wherein the similarity outcome is further based on a similarity outcome score of a number of text-based associations greater than or equal to the text-based threshold confidence level and a number of semantic-based associations greater than or equal to the semantic-based threshold confidence level.

18. The system of claim 17 , wherein the heuristic agent is further configured to:

evaluate at least one of: the candidate selection rule or the similarity rule, said evaluation based on a plurality of sorted similarity outcome scores associated with a plurality of pairings of ontologies; and

select at least one of: the candidate selection rule from a plurality of candidate selection rules or the similarity rule from a plurality of similarity rules based on said evaluation.

19. A method for ontology candidate selection and comparison comprising:

in a microprocessor, generating a set of ontology combinations from a plurality of ontologies, each ontology combination including a first ontology candidate and second ontology candidate;

comparing the first ontology candidate and the second ontology candidate in each ontology combination based upon a candidate selection rule to generate a plurality of comparisons of the ontology combinations;

based upon said comparisons, selecting a pair of ontologies from the plurality of ontology candidates;

generating a similarity outcome related to the pair of ontologies based upon a similarity rule; and

evaluating at least one of: the candidate selection rule or the similarity rule based upon the similarity outcome,

wherein the pair of ontologies includes a first ontology and a second ontology and the similarity rule is based upon:

a plurality of text-based confidence levels associated with a plurality of a text-based associations, each text-based association including concept strings from the first ontology and the second ontology;

a text-based threshold confidence level;

a plurality of semantic confidence levels associated with a plurality of semantic-based associations, each semantic-based association including concept strings from the first ontology and the second ontology; and

a semantic-based threshold confidence level, and the similarity outcome is based upon:

at least one semantic-based association associated with a semantic-based confidence level greater than or equal to the semantic-based threshold confidence level; and

at least one text-based association associated with a text-based confidence level greater than or equal to the text-based threshold confidence level.

20. The method of claim 19 , wherein said selecting from the plurality of ontology candidates a pair of ontologies further comprises:

sorting the plurality of comparisons; and

selecting the pair of ontologies according to the sorted comparisons.

21. The method of claim 19 , wherein the candidate selection rule is associated with at least one of: ontology observational criteria, ontology structural criteria, ontology functional criteria, ontology processing time criteria, or ontology usability criteria.

22. The method of claim 21 , wherein said selecting from the plurality of ontology candidates a pair of ontologies further comprises:

selecting the pair of ontologies based on a comparison of at least one of: an ontology structural criterion of the first ontology and an ontology structural criterion of the second ontology or an ontology functional criterion of the first ontology and an ontology functional criterion of the second ontology.

23. The method of claim 19 , wherein the pair of ontologies includes a first ontology and a second ontology and the similarity rule is based on a confidence level associated with a text-based association including a concept string from the first ontology and a concept string from the second ontology.

24. The method of claim 23 , wherein the similarity rule is further based on a threshold confidence level and the similarity outcome is based on at least one text-based association associated with a confidence level greater than or equal to the threshold confidence level.

25. The method of claim 24 , wherein said generating a similarity outcome related to the pair of ontologies further comprises:

generating the threshold confidence level based on an F-score associated with the at least one text-based association.

26. The method of claim 19 , wherein the pair of ontologies includes a first ontology and a second ontology and the similarity rule is based on a confidence level associated with a semantic-based association including a concept from the first ontology and a concept from the second ontology.

27. The method of claim 26 , wherein the similarity rule is further based on a threshold confidence level and the similarity outcome includes at least one semantic-based association associated with a confidence level greater than or equal to the threshold confidence level.

28. The method of claim 27 , wherein said generating a similarity outcome related to the pair of ontologies further comprises:

generating the threshold confidence level based on an F-score and a significant difference calculation associated with the at least one semantic-based association.

29. The method of claim 19 , wherein the similarity outcome is further based on a similarity outcome score of a number of text-based associations greater than or equal to the text-based threshold confidence level and a number of semantic-based associations greater than or equal to the semantic-based threshold confidence level.

30. The method of claim 29 , wherein said evaluating at least one of: the candidate selection rule or the similarity rule is further based on a plurality of sorted similarity outcome scores associated with a plurality of ontology pairings.

31. The method of claim 19 , wherein said evaluating at least one of: the candidate selection rule or the similarity rule is further based on a plurality of sorted similarity outcomes associated with a plurality of ontology pairings.

32. The method of claim 19 , further comprising:

generating an alignment mapping between similar concepts in the pair of ontologies based on the similarity outcome.

33. The method of claim 32 , wherein said generating an alignment mapping further comprises:

generating a new or modified ontology, schema, or relationship list based on the alignment mapping.

Assignments (12)

Continuity (1)