Bayesian Network Based Expert System for Seismic Performance Assessment

摘要

In this paper, Bayesian network based expert system is built for seismic performance evaluation. Unlike the regular rule-based expert system which allows only unidirectional reasoning with strict complete data entry sets, the Bayesian network based expert system can perform bidirectional inference based on incomplete information when the prior probability distributions are given. Seismic performance evaluation is a task that often needs experiences to draw a conclusion under insufficient information circumstance, which made the Bayesian network- based expert system well qualified for the job, and the example also illustrates this point.