Vocabulary Model Requirements for Production Rule Systems

Software solutions for certain complex problem areas such as claims adjudication and fraud detection are hard to build using imperative programming languages and even harder to maintain and change. In light of this, enterprises are turning to Rule-Based Expert Systems to provide a solution. However, to realize the full potential of this, the technology for mining and articulating the knowledge need to be made available to the Subject Matter Experts in a natural language like format. Further, this knowledge is best captured as production rules that can preferably be transformed and executed on production rule engines. Semantics of Business Vocabulary and Business Rules (SBVR) specification provides a framework for defining a formal language for specifying high level non-production business rules. The intent of SBVR is to capture these high level business rules in a formalized unambiguous format. However, this does not readily lend itself to a Production Rule System format. This paper explores the requirements surrounding the modeling of domain vocabularies that can be used to specify user friendly rules which can readily be transformed to executable production rule syntax and semantics of a target production rule engine.