Tilbake til søkeresultatene

VERDIKT-Kjernekomp.og verdiskaping IKT

Cooperative Mining of Independent Document Repositories

Tildelt: kr 8,6 mill.

Document repositories contain a wealth of knowledge that can be discovered through text-mining techniques. With the advent of all-digital management processes, new types of repositories have emerged, for example personal document repositories like email folders. The typical approach to mine collections from several repositories has been to logically merge the collections before performing the mining process. However, for many application areas this is not possible. For example, some collections can not b e merged for legal reasons, while others can not be merged because of risk of revealing classified information. In addition, they might be based on the use of different schemas or ontologies, making merging non-trivial. The general approach to the proble m is independent processing of collections at each repository, creating sufficient intermediate data to perform global mining. However, it is not obvious what intermediate data is needed, and use of different schemas and semantics means the process also needs some kind of schema/ontology mediation. In the COMIDOR project we aim at solving those challenges, through development of the mechanisms and algorithms necessary to succeed in the process of cooperative mining of independent document repositories .

Budsjettformål:

VERDIKT-Kjernekomp.og verdiskaping IKT