Modelling Second-Order Uncertainty in State Machines

Modelling the behaviour of state-based systems can be challenging, especially when modeller is not entirely certain about its intended interactions with user or environment. Currently, it possible to associate a stated level uncertainty given event by attaching probabilities transitions (producing ‘Probabilistic State Machines’). This captures ‘First-order uncertainty’ - (un-)certainty that will occur. However, this does permit capture their own (or lack thereof) probability also known as ‘Second-order uncertainty’. In article we introduce generalisation probabilistic finite state machines makes incorporate important additional dimension uncertainty. For adopt formalism for reasoning called Subjective Logic. We present an algorithm create these enhanced automatically from conventional machine and set observed sequences. show how approach used reverse-engineering predictive traces.