RobSSM: R Code for Robust Estimation for State Space Models

These separate R files provide functions to simulate data according to the North Sea pollock stock assessment state space model (SSM) and fit the model according to (Laplace-approximated) maximum likelihood (ML) and robust estimation methods. The RobSSM_Main.r script guides through all functions and generates plots similar to those in the paper. Estimation relies on the R package Template Model Builder (TMB). Details about the robust methodology and theoretical properties can be found in Aeberhard et al. (2020).

Updates can be found at https://github.com/williamaeberhard/robssm.

Any requests/comments/bug reports should be sent to william.aeberhard@gmail.com.

Contents

Files contained in this repository:

• RobSSM_Main.r: main R file that goes through all functions, simulates data, contaminates them, and compares ML and robust outputs;
• NP_nst_gen.r: generates data according to the non-stationary (nst) SSM used for the assessment of pollock in the North Sea;
• NP_nst_nrcorrect.r: Newton-Raphson correction of the robustified gradient by Monte Carlo approximation of the expected gradient;
• NP_nst_transfo.r: transformation of model parameters, to feed optimization;
• NP_nst_untransfo.r: un-transformation of model parameters, after optimization;
• NP_nst_weights.r: robustness weights according to smooth semi-Huber or log-logistic rho function;
• NP_nst.cpp: C++ TMP template for nst SSM, for both ML and robust estimation;
• NP_st_gen.r: generate data according to the stationary (st) SSM used for the assessment of pollock in the North Sea;
• NP_st_nrcorrect.r: Newton-Raphson correction of the robustified gradient by Monte Carlo approximation of the expected gradient;
• NP_st_stdist.r: computes parameters (means and variances) of stationary distribution, used for initial conditions;
• NP_st_transfo.r: transformation of model parameters, to feed optimization;
• NP_st_untransfo.r: un-transformatio of model parameters, after optimization;
• NP_st_weights.r: robustness weights according to smooth semi-Huber or log-logistic rho function;
• NP_st.cpp: C++ TMP template for st SSM, for both ML and robust estimation;
• this README file.

Version History

This is RobSSM version 0.1. This is the initial release.

References

Aeberhard, W. H., Cantoni, E., Field, C. Künsch, H. R., Mills Flemming, J., and Xu, X. (2020) Robust Estimation for Discrete-Time State Space Models. Scandinavian Journal of Statistics, In Press. DOI: 10.1111/sjos.12482. Preprint: https://arxiv.org/abs/2004.05023