BAYESIAN THRESHOLD AUTOREGRESSIVE MODELS FOR NONLINEAR TIME SERIES |
| |
| Authors: | John Geweke Nobuhiko Terui |
| |
| Affiliation: | Federal Reserve Bank of Minneapolis and University of Minnesota, and University of Minnesota and Yamagata University |
| |
| Abstract: | Abstract. This paper provides a Bayesian approach to statistical inference in the threshold autoregressive model for time series. The exact posterior distribution of the delay and threshold parameters is derived, as is the multi-step-ahead predictive density. The proposed methods are applied to the Wolfe's sunspot and Canadian lynx data sets. |
| |
| Keywords: | Nonlinear time series threshold model Monte Carlo integration regime change prediction Wolfe's sunspot Canadian lynx |
|
