Improved Bayesian Network-Based Risk Model and Its Application in Disaster Risk Assessment

• DOI: 10.1007/s13753-018-0171-z
• Publication Date: 2018-05-15T11:34:48Z
• Authors: Ming Li; Mei Hong; Ren Zhang
• Keywords: Decorrelation

The application of Bayesian network (BN) theory in risk assessment is an emerging trend. But cases where data are incomplete and variables mutually related, its restricted. To overcome these problems, improved BN model with parameter retrieval decorrelation ability proposed. First, multivariate nonlinear planning applied to the feedback error learning parameters. A genetic algorithm used learn probability distribution nodes that lack quantitative data. Then, based on grey relational analysis considers correlation variation rate, optimal weight characterizes calculated weighted for decorrelation. An then built. sea ice disaster shows can be variable correlation.