Parallel deterministic procedure based on hidden Markov models for the analysis of economic cycles in Poland

In the paper the deterministic version of the procedure based on hidden Markov models for the analysis of economic cycles is described. The quality of fitting hidden Markov models as well as the accuracy of the identification of turning points in the business cycle in Poland depends, among other things, on the number of states of the model and the size of panel data. Determinism however affects significantly on a time of computations. Speed up of computations could be achieved by adding the parallelism into the procedure. The usefulness of this approach is verified by the numerical experiments and comparative tests measuring a time of computations depending on the number of processor cores.