10.7 Interpreting the MARSS output

By now the MARSS() output should look familiar. The first 12 parameter estimates Z.z## are the loadings of each observed time series on the 3 hidden states. The next 5 estimates R.(,) are the variances of the observation errors \((v_{i,t})\). The last 3 values, x0.X#, are the estimates of the initial states at \(t = 0\).

Recall that the estimates of the processes themselves (i.e., \(\mathbf{x})\) are contained in one of the list elements in our fitted MARSS object. Specifically, they are in mod_fit$states, and their respective standard errors are in mod_fit$states.se. For the names of all of the other objects, type names(dfa_1).