METHOD FOR FORECASTING UNSTABLE POLICY ENFORCEMENTS

摘要

Method is for predicting that non-stable strategy is implemented, description, wherein behavior dynamic bayesian network (DBN) model and tactful finite state transducer is extended is equipped with tautness function and identity (TFFST) model is analysis composition, derive the system that the given strategy of the implementation of forecasting consequence, especially flip cover clap configuration change. The method comprising the steps of: that holding behavior model (BM) of-translation (1) Bayesian network is equipped with tautness function and identity (TFFST) to finite state transducer is extended; Joint Bayesian network (BM) and Policy model (PM) finite state transducer for calculating (2) are extended equipped with tautness function and identity (TFFSTs); It constitutes in the extended previous step generated equipped with tautness function and identity (TFFST) of (3) finite state transducer in itself; And each possible path output and input of-detection (4,5,6,7) repeated events,; If at least one repeats to be found, possible unstability (9) are detected.