Integrate topic categorizer to be implemented by Stanbol
--------------------------------------------------------
Key: NXSEM-44
URL: https://jira.nuxeo.com/browse/NXSEM-44
Project: Nuxeo Semantic R&D
Issue Type: New Feature
Reporter: Olivier Grisel
Assignee: Olivier Grisel
Integrate the 50000 topics categorizer prototyped for IKS early adopters
meeting in July.
https://issues.apache.org/jira/browse/STANBOL-197
Once this is implemented on the Stanbol side, the Nuxeo connector should be
leveraged to classify documents according to their top 3 topics (in addition to
extracting occurrences of entities).
E.g. a document can be classified as talking about "Economy Crisis", "European
Economy" and "French Presidential Elections" as document topics while having
explicit mentions of "Nicolas Sarkozy", "Francois Holland" as entities of type
Person and "Paris" and "France" as entities of type "Place".
This would probably imply to refactor the way we handle the types for entities:
use dynamic facets and have a base type called "SemanticResource" with subtypes
"Concept" and "Entity".
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
_______________________________________________
ECM-tickets mailing list
[email protected]
http://lists.nuxeo.com/mailman/listinfo/ecm-tickets
