A new approach for computing conditional probabilities of general stochastic processes

In this paper, hidden Markov model algorithms are considered as a method for computing conditional properties of continuous-time stochastic simulation models. The goal is to develop an algorithm that adapts known hidden Markov model algorithms for use with proxel-based simulation. It is shown how the forward- and Viterbi-algorithms can be directly integrated in the proxel-method. The possibility of integrating the more complex Baum-Welch-algorithm is theoretically addressed. Experiments are conducted to determine the practicability of the new approach and to illustrate the type of analysis that is possible.