Methods of defining ontologies, word disambiguation methods, computer systems, and articles of manufacture

• US 8,036,876 B2
• Filed: 11/04/2005
• Issued: 10/11/2011
• Est. Priority Date: 11/04/2005
• Status: Active Grant

• Current Assignee: Battelle Memorial Institute
• Original Assignee: Battelle Memorial Institute
• Inventors: Posse, Christian, Gregory, Michelle L., Baddeley, Robert L., Sanfilippo, Antonio P., Whitney, Paul D., Tratz, Stephen C., Hohimer, Ryan E., Chappell, Alan R.
• Primary Examiner(s): Wozniak; James S.
• Assistant Examiner(s): GUERRA-ERAZO, EDGAR X
• Application Number:
  US11/267,370
• Publication Number:
  US 20070106493A1
• Time in Patent Office: 2,167 Days
• Field of Search: 704/3, 704/4, 704/7, 704/9, 704/10, 706/60, 715/260, 707/707
• US Class Current: 704/9
• CPC Class Codes: G06F 16/367 Ontology
• CPC Class Codes: G06F 40/30 Semantic analysis

Abstract

Methods of defining ontologies, word disambiguation methods, computer systems, and articles of manufacture are described according to some aspects. In one aspect, a word disambiguation method includes accessing textual content to be disambiguated, wherein the textual content comprises a plurality of words individually comprising a plurality of word senses, for an individual word of the textual content, identifying one of the word senses of the word as indicative of the meaning of the word in the textual content, for the individual word, selecting one of a plurality of event classes of a lexical database ontology using the identified word sense of the individual word, and for the individual word, associating the selected one of the event classes with the textual content to provide disambiguation of a meaning of the individual word in the textual content.

Claims

1. A computer-implemented method of defining a lexical database ontology comprising:

accessing a lexical database comprising a plurality of concepts individually including a plurality of different words;

accessing a corpus of documents;

using a processor, identifying occurrences of the words in the documents of the corpus;

using the processor, counting a plurality of frequency counts for the concepts, wherein the counting includes incrementing the frequency counts of the concepts using the identifying of the occurrences of the words of the respective concepts in the documents;

using the processor, analyzing the frequency counts;

using the processor, selecting a plurality of the concepts as event classes of a lexical database ontology using the analyzing, wherein the event classes of the lexical database ontology are usable to disambiguate textual content; and

wherein each concept includes a sense of a word lemma, and wherein the counting for one of the concepts comprises incrementing the frequency count of the one concept, for each occurrence of the word lemma in the documents of the corpus which is included in the one concept, by 1/n where n is the total number of concepts which include a sense of the world lemma.

2. The method of claim 1 wherein the concepts individually comprise a synonym set comprising a plurality of synonyms related to the individual concept.

3. The method of claim 1 wherein the concepts individually include the respective different words comprising verbs.

4. The method of claim 1 wherein the concepts individually include a plurality of senses for respective ones of the different words of the individual concept.

5. The method of claim 1 wherein the analyzing comprises comparing the frequency counts with respect to a threshold and the selecting comprises selecting the concepts whose frequency counts exceed the threshold.

6. The method of claim 1 wherein the incrementing comprises incrementing the frequency counts of concepts which are hypernyms of concepts whose frequency counts have been incremented.

7. The method of claim 1 wherein the selecting comprises selecting only concepts which include words which comprise hyponyms.

8. The method of claim 1 further comprising defining instances of the event classes, wherein the defining comprises for an individual one of the concepts, defining the instances to comprise words of the concepts and troponyms of the words of the concepts for the individual one of the concepts.

9. The method of claim 8 wherein the instances comprising words additionally comprise the senses of the words.

10. The method of claim 1 further comprising, for an individual one of the event classes, selecting one of the words of the respective concept of the event class as the name of the respective event class for use in disambiguation of instances of the event class.

11. The method of claim 10 wherein the selecting the one of the words comprises selecting the one of the words having the highest frequency of occurrence in the documents.

12. The method of claim 10 wherein the accessing the lexical database comprises accessing a version of WordNet.

17. The method of claim 1 wherein the event classes of the lexical database ontology are usable to disambiguate the textual content of text to be analyzed subsequently after the defining the lexical database ontology including the selecting the concepts as event classes of the lexical database ontology.

13. An article of manufacture comprising:

non-transistory media storing programming configured to cause processing circuitry to perform processing comprising;

accessing a lexical database comprising a plurality of concepts individually including a plurality of different words;

selecting a plurality of the concepts as event classes of a lexical database ontology, wherein the event classes of the lexical database ontology are usable to disambiguate textual content;

accessing a corpus of documents;

identifying occurrences of the words in the documents of the corpus;

counting a plurality of frequency counts for the concepts, wherein the counting includes incrementing the frequency counts of the concepts as a result of the identifying of the occurrences of the words of the respective concepts in the documents;

analyzing the frequency counts, and wherein the selecting comprises selecting using the analyzing; and

wherein each concept includes a sense of a word lemma, and wherein the counting for one of the concepts comprises incrementing the frequency count of the one concept, for each occurrence of the word lemma in the documents of the corpus which is included in the one concept, by 1/n where n is the total number of concepts which include a sense of the world lemma.

14. The article of claim 13 wherein the processing further comprises defining instances of the event classes, wherein the defining comprises for an individual one of the concepts, defining the instances to comprise words of the concepts and troponyms of the words of the concepts for the individual one of the concepts.

15. The article of claim 14 wherein the instances comprising words additionally comprise the senses of the words.

16. The article of claim 13 wherein the processing further comprises, for an individual one of the event classes, selecting one of the words of the respective concept of the event class as the name of the respective event class for use in disambiguation of instances of the event class.

18. The article of claim 13 wherein the textual content to be disambiguated comprises a plurality of text items which individually comprise a plurality of words.