Knowledge acquisition is a key issue in the business rule methodology. As Natural Language (NL) policies and regulations are often important or even contractual sources of knowledge, we propose a framework for acquisition and maintenance of business rules based on NL texts. It enables business experts to author the specification of rule applications without the help of knowledge engineers. This framework has been created as part of the ONTORULE project, which is defining an integrated platform for acquisition, maintenance and execution of business-oriented knowledge bases combining ontologies and rules.
Our framework relies on a data structure, called ”index”, encompassing and connecting the source text, the ontology and a textual representation of rules. Textual rules are as close to the Structured English representation of SBVR as possible for business users in charge of rule elicitation. The index relies on W3C technologies, which makes the tools interoperable and enable semantic search. We show that such an index structure supports the parallel maintenance of policy documents and knowledge bases (acquisition, consistency check and update).
Two detailed examples with preliminary results are provided, one from air travel and the other from the automotive industry.