Destagionalizzazione di serie attuariali mediante un modello Bayesiano

D. Spelta
In this paper we consider a Bayesian model of decomposition of time series proposed by Akaike. The computational cost of the algorithm has been reduced by Corradi and Scarani, who considered only the case of quarterly time series. Here the model has been made more flexible and the algorithm has been applied to time series of actuarial interest.