| Abstract |
This talk presents some recent work on the Tmeko procedure that will guide non-expert in creating complex mapping for domain wordnet synsets to a central ontology. The procedure was designed to meet the requirements of the Kyoto project that involves creating a wiki-like platform and a knowledge base for information sharing for users in the environment domain.
One of the central notions of the procedure involves mining and processing definitions in order to find the ontological classes the synset needs to be related to. The users will validate these candidate classes and the type of relation by answering simple natural language questions. Some future versions and use cases of the system will be discussed, especially a (semi-) automatic version that does not depend on user participation.
|