Chapter 16 Algorithm notes and cautions

The MARSS package fits models via maximum likelihood. The MARSS package is unusual among packages for fitting MARSS models in that fitting is performed via a constrained EM algorithm (Holmes 2013) based on a vectorized form of Equation (1.1).

Although fitting via the BFGS algorithm is also provided using method="BFGS" and the optim function in R, the examples in this guide use the EM algorithm primarily because it gives robust estimation for datasets replete with missing values and for high-dimensional models with various constraints. However, there are many models/datasets where BFGS is faster and we typically try both for problems. All the EM code in the MARSS package is currently in native R. Thus the model fitting with the EM algorithm is rather slow.