Mutual information model selection algorithm for time series

The research article “Mutual information model selection algorithm for time series”, co-authored by METU member Assoc. Prof. Ceylan Yozgatlıgil, has been published in Journal of Applied Statistics.

Time series model selection has been widely studied in recent years. It is of importance to select the best model among candidate models proposed for a series in terms of explaining the procedure that governs the series and providing the most accurate forecast for the future observations. In this study, it is aimed to create an algorithm for order selection in Box–Jenkins models that combines penalized natural logarithm of mutual information among the original series and predictions coming from each candidate. The penalization is achieved by subtracting the number of parameters in each candidate and empirical information the data provide.Simulation studies under various scenarios and applications on real data sets imply that our algorithm offers a promising and satisfactory alternative to its counterparts.

Akça, E., & Yozgatlıgil, C. (2020). Mutual information model selection algorithm for time series. Journal of Applied Statistics, 47(12), 2192-2207. doi:10.1080/02664763.2019.1707516

Article access: https://www.tandfonline.com/doi/full/10.1080/02664763.2019.1707516

Keywords:
Box–Jenkins models; mutual information; order selection

Other authors:
Akça, E.

Acknowledgments:
This study was supported by BAP research grant of Faculty of Arts and Sciences of Middle East Technical University (project number: BAP-01-09-2017-001).