Threat assessment based on variable parameter dynamic Bayesian network

To resolve the key issue of uncertainty knowledge representation in the process of establishing threat assessment system, a variable-parameter dynamic Bayesian network for threat assessment is proposed. A self-adaptive assessment system based on variable-parameter dynamic Bayesian network is presented. And a novel threat assessment optimization algorithm based on variable-parameter dynamic Bayesian network is developed, which achieves the goal of learning variable parameters by on-line data mining techniques. The experimental results demonstrate that, using real-time data, the presented approach can modify the threat assessment knowledge repository dynamically, which leads the assessment model to a better adaptability for producing more accurate assessment results.