Prediction of a complex system with few data: Evaluation of the effect of model structure and amount of data with dynamic bayesian network models

Author(s): Maldonado, A.D., Uusitalo, L., Tucker, A., Blenckner, T., Aguilera, P.A., Salmerón, A.
In: Environmental Modelling & SoftwareVolume 118, August 2019, Pages 281-297
Year: 2019
Type: Journal / article
Theme affiliation: Marine
Link to centre authors: Blenckner, Thorsten
Full reference: Maldonado, A.D., Uusitalo, L., Tucker, A., Blenckner, T., Aguilera, P.A., Salmerón, A. 2019. Prediction of a complex system with few data: Evaluation of the effect of model structure and amount of data with dynamic bayesian network models. Environmental Modelling & Software Volume 118, August 2019, Pages 281-297

Summary

A major challenge in environmental modeling is to identify structural changes in the ecosystem across time, i.e., changes in the underlying process that generates the data.

In this paper, we analyze the Baltic Sea food web in order to 1) examine potential unobserved processes that could affect the ecosystem and 2) make predictions on some variables of interest. To do so, dynamic Bayesian networks with different setups of hidden variables (HVs) were built and validated applying two techniques: rolling-origin and rolling-window. Moreover, two statistical inference approaches were compared at regime shift detection: fully Bayesian and Maximum Likelihood Estimation.

Our results confirm that, from the predictive accuracy point of view, more data help to improve the predictions whereas the different setups of HVs did not make a critical difference in the predictions. Finally, the different HVs picked up patterns in the data, which revealed changes in different parts of the ecosystem.

Share

Stockholm Resilience Centre is a collaboration between Stockholm University and the Beijer Institute of Ecological Economics at the Royal Swedish Academy of Sciences

Stockholm Resilience Centre
Stockholm University, Kräftriket 2B
SE-10691
Phone: +46 8 674 70 70
info@stockholmresilience.su.se

Organisation number: 202100-3062
VAT No: SE202100306201