In this project we study problems of inference and forecasting in autoregressive models with periodic correlation from a Bayesian perspective. Normality and unimodality assumptions are rarely verified in practice and the usual approach is to try Box-Cox transformations to obtain approximate normality and stabilize the periodic variance. More recently, mixture models were developed to take into account asymmetry and multimodality.

1. Ricardo Ehlers (UFPR)
2. Marinho Gomes de Andrade (USP)

[2002, article]

Lewis, P. A. W., & Ray, B. K. (2002). Nonlinear Modelling of Periodic Threshold Autoregressions using TSMARS. Journal of Time Series Analysis, 23(4), 459-471.

[2006, article]

Shao, Q. (2006). Mixture Periodic Autoregressive Time Series Models. Statistics and Probability Letters, 76, 609-618.