![]() | Only 14 pages are availabe for public view |
Abstract Multiple seasonal patterns are noticeable in time series data. Therefore, seasonal autoregressive moving average (SARMA) models have been recently extended to double SARMA (DSARMA) models. In this study, DSARMA models is extended to double seasonal autoregressive moving average- generalized autoregressive conditional heteroskedasticity (DSARMA-GARCH) in order not only to capture multiple seasonal patterns but also to take into account the volatility of the series at the same time. A Bayesian approach is used here to estimate these models. Although, DSARMA-GARCH models are non-linear in their coefficients, the Metropolis-Hastings (MH) algorithm is one of the most used Markov Chain Monte Carlo (MCMC) methods to overcome this problem.Therefore, the MH algorithm is used and investigated to provide Bayesian estimation of DSARMA-GARCH models. The obtained results demonstrate that this algorithm is suitable for Bayesian estimation of DSARMA-GARCH models |