CAPTURING INFORMATION ON BEHAVIOR WITH THE RADD-NLI - A LINGUISTIC AND KNOWLEDGE-BASED APPROACH

In this paper we illustrate how information on behaviour can be captur ed from natural language with the RADD-NLI. The behavioural informatio n contained in verbal descriptions of processes is transformed into th e RADD-NLI process model of characteristic processes. This process mod el is then integrated into the knowledge base of the RADD-NLI. Althoug h taken from a specific domain this model is abstract enough to form t he basis for further applications with synonym relations. Processes co ming up in the user dialogue will be matched against the knowledge bas e. A dialogue control is activated if a mismatch occurs or for a detai led discussion of captured processes with the user. Linguistic knowled ge, especially antonymous verb pairs, is considered to decide whether a captured process will be integrated into the structural database des ign and/or the process list of the design.