Evaluation of the SRA tool using Data Mining Techniques

Abstract

This paper describes a validation approach of a socio-technical design support system using data mining techniques. Bayesian Belief Networks (BBN) are used to assess human error and system failure [13] based on a variety of high-level operational scenarios. The System Reliability Analyser (SRA) tool automates the process by iteratively manipulating the BBN model. Data mining techniques are employed in order to identify whether the initial assumptions embedded in the system reliability model are met by results from scenario-based testing.