From Kolflow Project

Revision as of 07:43, 5 February 2011 by Desmontils-e (Talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Refering task : Task2

Title : Formal methods for knowledge extraction.

Code : 2.2

Start Date :

End Date :

Description : This task aims at defining a continuous process of knowledge extraction which uses recent works we carried out in Orpailleur on knowledge extraction. Indeed, Bendaoud’s thesis[18] uses formal concept analysis to build a conceptualisation of the domain from raw texts. After validation by experts, this conceptual- isation is transformed in an ontology encoded in description logics. A prototype implements this methodology that was tested in astronomy, in microbiology and some attempts in the Taaable project for building a cook ontology [16]. The strength of such a formal approach is that building an ontology is now a process guided by the corpus less time consuming than a full-manual process. Subjectivity in building the ontology is also reduced with such a process. However, this work suffers from several weaknesses that the following subtasks T2.3 and T2.4 will manage. T2.2 aims at integrating our prototype for building a conceptualisation from texts into a semantic wiki environment. This requires that texts are entered in the semantic wiki as resources. Natural language processing as well as formal concept analysis (FCA) will be defined as modules. And finally, results from FCA – the conceptualisation – should be encoded as the ontology in the semantic wiki. Results of this tasks – methodology, modules, and experiments – will be recorded in deliverable D2.2.

Deliverables :

description Feb2011+months
D22 Integrating a knowledge extraction system from texts in a semantic wiki. methodology, modules, and experiments. 1212