Repository logo
 

Modeling directional (circular) time series


Type

Working Paper

Change log

Authors

Harvey, A. 
Hurn, S. 
Thiele, S. 

Abstract

Circular observations pose special problems for time series modeling. This article shows how the score-driven approach, developed primarily in econometrics, provides a natural solution to the difficulties and leads to a coherent and unified methodology for estimation, model selection and testing. The new methods are illustrated with hourly data on wind direction.

Description

Keywords

Autoregression, circular data, dynamic conditional score model, von Mises distribution, wind direction

Is Part Of

Publisher

Faculty of Economics

Publisher DOI

Publisher URL