Initialize.reStruct {nlme} | R Documentation |

## Initialize reStruct Object

### Description

Initial estimates for the parameters in the `pdMat`

objects
forming `object`

, which have not yet been initialized, are
obtained using the methodology described in Bates and Pinheiro
(1998). These estimates may be refined using a series of EM
iterations, as described in Bates and Pinheiro (1998). The number of
EM iterations to be used is defined in `control`

.

### Usage

## S3 method for class 'reStruct':
Initialize(object, data, conLin, control, ...)

### Arguments

`object` |
an object inheriting from class `reStruct` ,
representing a random effects structure and consisting of a list of
`pdMat` objects. |

`data` |
a data frame in which to evaluate the variables defined in
`formula(object)` . |

`conLin` |
a condensed linear model object, consisting of a list
with components `"Xy"` , corresponding to a regression matrix
(`X` ) combined with a response vector (`y` ), and
`"logLik"` , corresponding to the log-likelihood of the
underlying model. |

`control` |
an optional list with a single component `niterEM`
controlling the number of iterations for the EM algorithm used to
refine initial parameter estimates. It is given as a list for
compatibility with other `Initialize` methods. Defaults to
`list(niterEM = 20)` . |

`...` |
some methods for this generic require additional
arguments. None are used in this method. |

### Value

an `reStruct`

object similar to `object`

, but with all
`pdMat`

components initialized.

### Author(s)

Jose Pinheiro Jose.Pinheiro@pharma.novartis.com and Douglas Bates bates@stat.wisc.edu

### References

Bates, D.M. and Pinheiro, J.C. (1998) "Computational methods for
multilevel models" available in PostScript or PDF formats at
http://franz.stat.wisc.edu/pub/NLME/

### See Also

`reStruct`

, `pdMat`

[Package

*nlme* version 3.1-66

Index]