mxRestore {OpenMx}

Restore model state from a checkpoint file

Description

Restore model state from a checkpoint file

Usage

mxRestore(
  model,
  chkpt.directory = mxOption(model, "Checkpoint directory"),
  chkpt.prefix = mxOption(model, "Checkpoint Prefix"),
  line = NULL,
  strict = FALSE
)

mxRestoreFromDataFrame(model, checkpoint, line = NULL)

Arguments

model

an MxModel object

chkpt.directory

character. Directory where the checkpoint file is located

chkpt.prefix

character. Prefix of the checkpoint file

line

integer. Which line from the checkpoint file to restore (defaults to the last line)

strict

logical. Require that the checkpoint name and model name match

checkpoint

a data.frame containing the model state

Details

In general, the arguments ‘chkpt.directory’ and ‘chkpt.prefix’ should be identical to the mxOption: ‘Checkpoint Directory’ and ‘Checkpoint Prefix’ that were specified on the model before execution.

Alternatively, the checkpoint file can be manually loaded as a data.frame in R and passed to mxRestoreFromDataFrame. Use read.table with the options header=TRUE, sep="\t", stringsAsFactors=FALSE, check.names=FALSE.

Value

Returns an MxModel object with free parameters updated to the last saved values. When ‘line’ is provided, the MxModel is updated to the values on that line within the checkpoint file.

References

The OpenMx User's guide can be found at https://openmx.ssri.psu.edu/documentation

See Also

Other model state: mxComputeCheckpoint(), mxSave()

Examples

library(OpenMx)

# Simulate some data

x=rnorm(1000, mean=0, sd=1)
y= 0.5*x + rnorm(1000, mean=0, sd=1)
tmpFrame <- data.frame(x, y)
tmpNames <- names(tmpFrame)

dir <- tempdir()  # safe place to create files
mxOption(key="Checkpoint Directory", value=dir)

# Create a model that includes an expected covariance matrix,
# an expectation function, a fit function, and an observed covariance matrix

data <- mxData(cov(tmpFrame), type="cov", numObs = 1000)
expCov <- mxMatrix(type="Symm", nrow=2, ncol=2, values=c(.2,.1,.2), free=TRUE, name="expCov")
expFunction <- mxExpectationNormal(covariance="expCov", dimnames=tmpNames)
fitFunction <- mxFitFunctionML()
testModel <- mxModel(model="testModel", expCov, data, expFunction, fitFunction)

#Use mxRun to optimize the free parameters in the expected covariance matrix
modelOut <- mxRun(testModel, checkpoint = TRUE)
modelOut$expCov

#Use mxRestore to load the last checkpoint saved state of the model
modelRestore <- mxRestore(testModel)
modelRestore$expCov