print.marssMLE {MARSS}

Printing functions for MARSS MLE objects

Description

MARSS() outputs marssMLE objects. print(MLEobj), where MLEobj is a marssMLE object, will print out information on the fit. However, print can be used to print a variety of information (residuals, smoothed states, imputed missing values, etc) from a marssMLE object using the what argument in the print call.

Usage

## S3 method for class 'marssMLE'
print(x, digits = max(3, getOption("digits")-4), ..., 
    what = "fit", form = NULL, silent = FALSE)

Arguments

x

A marssMLE object.

digits

Number of digits for printing.

...

Other arguments for print.

what

What to print. Default is "fit". If you input what as a vector, print returns a list. See examples.

"model"

The model parameters with names for the estimated parameters. The output is customized by the form of the model that was fit. This info is in attr(x$model, "form") .

"par"

A list of only the estimated values in each matrix. Each model matrix has it's own list element. Standard function: coef(x)

"start" or "inits"

The values that the optimization algorithm was started at. Note, x$start shows this in form="marss" while print shows it in whatever form is in attr(x$model, "form") .

"paramvector"

A vector of all the estimated values in each matrix. Standard function: coef(x, type="vector"). See coef().

"par.se","par.bias","par.lowCIs","par.upCIs"

A vector the estimated parameter standard errors, parameter bias, lower and upper confidence intervals. Standard function: MARSSparamCIs(x) See MARSSparamCIs().

"xtT" or "states"

The estimated states conditioned on all the data. x$states

"data"

The data. This is in x$model$data

"logLik"

The log-likelihood. Standard function: x$logLik. See MARSSkf() for a discussion of the computation of the log-likelihood for MARSS models.

"ytT"

The expected value of the data conditioned on all the data. Returns the data if present and the expected value if missing. This is in x$ytT (ytT is analogous to xtT).

"states.se"

The state standard errors. x$states.se

"states.cis"

Approximate confidence intervals for the states. See MARSSparamCIs().

"model.residuals"

The one-step ahead model residuals or innovations. \mathbf{y}_t - \textrm{E}[\mathbf{Y}_t|\mathbf{y}_1^{t-1}], aka actual data at time t minus the expected value of the data conditioned on the data from t=1 to t-1. Standard function: residuals(x, type="tt1") See MARSSresiduals() for a discussion of residuals in the context of MARSS models.

"state.residuals"

The smoothed state residuals. \mathbf{x}_t^T- \textrm{E}[\mathbf{X}_t|\mathbf{x}_{t-1}^T], aka the expected value of the states at time t conditioned on all the data minus the expected value of the states at time t conditioned on \mathbf{x}_{t-1}^T]. Standard function: residuals(x, type="tT") See MARSSresiduals().

parameter name

Returns the parameter matrix for that parameter with fixed values at their fixed values and the estimated values at their estimated values. Standard function: coef(x, type="matrix")$elem

"kfs"

The Kalman filter and smoother output. See MARSSkf() for a description of the output. The full kf output is not normally attached to the output from a MARSS() call. This will run the filter/smoother if needed and return the list invisibly. So assign the output as foo=print(x,what="kfs")

"Ey"

The expectations involving y conditioned on all the data. See MARSShatyt() for a discussion of these expectations. This output is not normally attached to the output from a MARSS() call–except ytT which is the predicted value of any missing y. The list is returned invisibly so assign the output as foo=print(x,what="Ey").

form

By default, print uses the model form specified in the call to MARSS(). This information is in attr(marssMLE$model, "form") , however you can specify a different form. form="marss" should always work since this is the model form in which the model objects are stored (in marssMLE$marss).

silent

If TRUE, do not print just return the object. If print call is assigned, nothing will be printed. See examples. If what="fit", there is always output printed.

Value

A print out of information. If you assign the print call to a value, then you can reference the output. See the examples.

Author(s)

Eli Holmes, NOAA, Seattle, USA.

Examples

 
  dat <- t(harborSeal)
  dat <- dat[c(2,11),]
  MLEobj <- MARSS(dat)
  
  print(MLEobj)
  
  print(MLEobj, what="model")
  
  print(MLEobj,what="par")
  
  #silent doesn't mean silent unless the print output is assigned
  print(MLEobj, what="paramvector", silent=TRUE)
  tmp <- print(MLEobj, what="paramvector", silent=TRUE)
  #silent means some info on what you are printing is shown whether
  #or not the print output is assigned
  print(MLEobj, what="paramvector", silent=FALSE)
  tmp <- print(MLEobj, what="paramvector", silent=FALSE)
  
  cis <- print(MLEobj, what="states.cis")
  cis$up95CI
  
  vars <- print(MLEobj, what=c("R","Q"))