DERIVING ONTOLOGY BASED ON LINGUISTICS AND COMMUNITY TAG CLOUDS

• US 20100030552A1
• Filed: 08/01/2008
• Published: 02/04/2010
• Est. Priority Date: 08/01/2008
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: George, David G., Chen, Feng-Wei, Medicke, John A.
• Granted Patent: US 8,359,191 B2
• Time in Patent Office: Days
• US Class Current: 704/9
• CPC Class Codes: G06F 16/367 Ontology

Abstract

In some embodiments, a method comprises receiving a tag cloud including tags that hyperlink to web content. The method can also comprise separating the tags into different linguistic categories, assigning a weight to each tag, and grouping the tags into clusters, wherein tags in a cluster are associated with a context. The method can also include determining one or more domains for the tag clusters, wherein a domain is a broadest class that defines one or more of the tags in a linguistic category, determining a hierarchy for the tags based on the weights of the tags, and identifying linguistic relationships between the tags. The method can also comprise determining properties associated with one or more of the tags and one or more of the domains, wherein the tag'"'"'s properties are determined using linguistic analysis and storing the tags, the hierarchies, the linguistic relationships, and the properties.

Claims

1. A method comprising:

receiving a tag cloud including tags that hyperlink to web content;

separating the tags into linguistic categories;

assigning a weight to each tag, wherein the weight is based on how many times the tag appears in the web content;

grouping the tags into clusters, wherein each tag in each cluster is associated with a context;

determining one or more domains for the tag clusters, wherein each domain defines one or more of the tags in each linguistic category;

determining a hierarchy for the tags based on the weights of the tags, wherein the hierarchy is based on results from a visual thesaurus;

identifying linguistic relationships between one or more of the tags;

determining properties associated with one or more of the tags and one or more of the domains, wherein the tag'"'"'s properties are determined using linguistic analysis.

2. The method of claim 1, wherein the linguistic categories include at least one of verbs, nouns, verbs, adjectives, synonyms, derived words, and translated words.

3. The method of claim 1, wherein the tag weights are assigned based on a ratio of how many times the tag appears in a website title associated with the web content plus how many times the tag links to the web content divided by how many times the tag appears in the web content.

4. The method of claim 1, wherein the determining the properties includes:

determining supplement information about the one or more tags, wherein the supplement information includes data from an online dictionary;

parsing the supplement information into a syntax structure tree, wherein the syntax structure tree includes parts of speech including noun phrases, verb phrases, and adjective phrases;

deriving ontology class attributes from the parts of speech.

5. The method of claim 1 further comprising:

determining one or more of the tags are associated with a domain that is part of an existing ontology;

determining one or more new attributes and relationships associated with the one or more of the tags and other tags within the domain;

augmenting the existing ontology to reflect the new attributes and relationships.

6. The method of claim 1 also comprising:

storing the tags, the hierarchies, the linguistic relationships, and the properties in an ontology repository;

receiving a request to generate a tag ontology, wherein the tag ontology is generated from information in the ontology repository;

determining relevant information from the ontology repository based on user-defined ontology parameters;

pruning the tag ontology based on a pruning threshold, wherein the pruning threshold determines a conciseness or verbosity of the tag ontology;

presenting the tag ontology, wherein the tag ontology can in the form of a hierarchy tree.

7. The method of claim 6, wherein tag ontology includes a classification of the one or more of the tags, an association of the tags with synonyms, an indication of context for the tags, and definitions of relationships between one or more of the tags.

8. The method of claim 6, wherein user-defined ontology parameters include one or more of pruning threshold, ontology class, ontology sub-class, ontology language, requirement of ontology attributes, ontology relationships, and ontology siblings.

9. The method of claim 6, wherein the tags with a weight lower than the pruning threshold are removed from the tag ontology.

10. The method of claim 6, further comprising:

dynamically generating the tag ontology in a user-specified ontology language.

11. A system comprising:

a tag cloud linguistic analyzer configured to receive a tag cloud including tags that hyperlink to web content, to separate the tags into linguistic categories, and to assign a weight to each tag, wherein the weight is based on how many times the tag appears in the web content;

a semantic domain analyzer configured to group the tags into clusters, wherein tags in each cluster are associated with a context, and to determine one or more domains for the tag clusters, wherein each domain defines one or more of the tags in one of the linguistic categories;

a taxonomy builder configured to determine a hierarchy for the tags based on the weights of the tags, wherein the hierarchy is based on results from a visual thesaurus;

a relationship analyzer configured to identify linguistic relationships between one or more of the tags;

an attribute analyzer configured to determine properties associated with one or more of the tags and one or more of the domains, wherein the tag'"'"'s properties are determined using linguistic analysis; and

an ontology repository to store the tags, the hierarchies, the linguistic relationships, and the properties.

12. The system of claim 11, wherein the different linguistic categories include at least one of verbs, nouns, verbs, adjectives, synonyms, derived words, and translated words.

13. The system of claim 11, wherein the tag weights are determined as a ratio of how many times each tag appears in a website title associated with the web content plus how many times the tag links to the web content divided by how many times the tag appears in the web content.

14. The system of claim 11, wherein the attribute analyzer is further configured

to determine supplement information about the one or more tags, wherein the supplement information includes data from an online dictionary,

to parse the supplement information into a syntax structure tree, wherein the syntax structure tree is used to determine different parts of speech including noun phrases, verb phrases, and adjective phrases, and

to derive ontology class attributes from the extracted parts of speech.

15. The system of claim 11, wherein the system is further configured to

determine one or more tags associated with a domain that is part of an existing ontology;

determine one or more new attributes and relationships associated with the one or more tags and other tags within the domain; and

augment the existing ontology to reflect the new attributes and relationships.

16. The system of claim 11 also comprising:

one or more applications configured to receive a request to generate a tag ontology, wherein the tag ontology is generated from information stored in an ontology repository;

an ontology generator configured to determine relevant information from the ontology repository based on user-defined ontology parameters and to present the tag ontology, wherein the tag ontology can in the form of a hierarchy tree, and

an ontology pruner configured to prune the tag ontology based on a pruning threshold, wherein the pruning threshold determines a conciseness or verbosity of the tag ontology.

17. The system of claim 16, wherein tag ontology includes a classification of the one or more of the tags, an association of the tags with synonyms, an indication of context for the tags, and definitions of relationships between one or more of the tags.

18. The system of claim 16, wherein user-defined ontology parameters include one or more of pruning threshold, ontology class, ontology sub-class, ontology language, requirement of ontology attributes, ontology relationships, and ontology siblings.

19. The system of claim 16, wherein the ontology pruner is configured to remove, from the tag ontology, tags with a weight lower than the pruning threshold.

20. The system of claim 16, wherein the ontology generator configured to dynamically generate in a user-specified ontology language including OWL and RDF, wherein an ontology language is used to encode ontology.

21. One or more machine-readable media having stored therein a program product, which when executed a set of one or more processor units causes the set of one or more processor units to perform operations comprising:

receiving a tag cloud including tags that hyperlink to web content;

separating the tags into linguistic categories;

assigning a weight to each tag, wherein the weight is based on how many times the tag appears in the web content;

grouping the tags into clusters, wherein each tag in each cluster is associated with a context;

determining one or more domains for the tag clusters, wherein each domain defines one or more of the tags in each linguistic category;

determining a hierarchy for the tags based on the weights of the tags, wherein the hierarchy is also based on results from a visual thesaurus;

identifying linguistic relationships between one or more of the tags;

determining properties associated with one or more of the tags and one or more of the domains, wherein the tag'"'"'s properties are determined using linguistic analysis.

22. The one or more machine readable media of claim 21, wherein the different linguistic categories include at least one of verbs, nouns, verbs, adjectives, synonyms, derived words, and translated words.

23. The one or more machine readable media of claim 21, wherein the weights are determined as a ratio of how many times the tag appears in a website title associated with the web content plus how many times the tag links to the web content divided by how many times the tag appears in the web content.

24. The one or more machine readable media of claim 21, wherein the determining the properties includes:

determining supplement information about the one or more tags, wherein the supplement information includes data from an online dictionary;

parsing the supplement information into a syntax structure tree, wherein the syntax structure tree is used to determine different parts of speech including noun phrases, verb phrases, and adjective phrases;

deriving ontology class attributes from the extracted parts of speech.

25. The one or more machine readable media of claim 21, further comprising:

determining one or more tags are associated with a domain that is part of an existing ontology;

determining one or more new attributes and relationships associated with the one or more tags and other tags within the domain;

augmenting the existing ontology to reflect the new attributes and relationships.