Keejam: a knowledge engineering methodology for expert judgment acquisition and modeling in probabilistic safety assessment

G. Cojazzi, G. Guida, L. Pinola, R. Sardella, P. Baroni

Proc. of ESREL 97 International Conference on Safety and Reliability, Lisbon, P, 1997, 199-207



In this paper a novel expert judgment methodology, called KEEJAM (Knowledge Engineering Expert Judgment Acquisition and Modeling), which is supposed to be applicable to a variety of expert judgment elicitation contexts, is described. The KEEJAM methodology takes in input: the definition of the application domain of interest and the specification of the expert judgment tasks to be faced. This approach provides structured and disciplined support to the normative expert, namely the knowledge engineer, in eliciting the knowledge and the reasoning strategies of the experts, building consistent knowledge models, and applying such models to the solution of the expert judgment tasks considered. The KEEJAM methodology is organized into five phases: start-up, design, knowledge acquisition and modeling, exploitation and refinement, synthesis and release. The main features of each phase are described in the paper, and practical suggestions for tailoring the KEEJAM methodology in concrete cases are provided. The KEEJAM methodology is currently being applied in the framework of a European Benchmark Exercise on Expert Judgment Techniques in Psa level 2.

Publisher's on-line resources

Return to publication list