SUPR
Locally stationary seasonal AR
Dnr:

NAISS 2023/22-1207

Type:

NAISS Small Compute

Principal Investigator:

Ganna Fagerberg

Affiliation:

Stockholms universitet

Start Date:

2023-11-27

End Date:

2024-12-01

Primary Classification:

10106: Probability Theory and Statistics

Allocation

Abstract

We propose a locally stationary seasonal AR process allowing for multiple periodicities and separate time-varying parameter processes in both the regular and seasonal parameters. Both the regular and seasonal parameters are parameterized to guarantee stationarity at every time point. The time evolution is modeled by dynamic shrinkage processes to allow for both longer periods without change and rapid jumps. A Gibbs sampler is developed with a particle Gibbs update step for the parameter trajectories. The model and the numerical effectiveness of the Gibbs sampler is investigated on simulate and real data.