mxJiggle {OpenMx}

Jiggle parameter values.

Description

Jiggle free parameter values, subject to box constraints. imxJiggle() is called internally by mxTryHard() (q.v.). mxJiggle() provides a more user-friendly wrapper to imxJiggle(), and can alternately emulate the 'JIGGLE' behavior of classic Mx.

Usage

mxJiggle(model, classic=FALSE, dsn=c("runif","rnorm","rcauchy"), loc=1, scale=0.25)
imxJiggle(params, lbounds, ubounds, dsn, loc, scale)

Arguments

model

An object of class MxModel.

classic

Logical; should mxJiggle() emulate the classic-Mx behavior elicited by keyword JIGGLE? Defaults to FALSE. See below, under "Details," for additional information.

dsn

Character string naming which random-number distribution–either uniform (rectangular), normal (Gaussian), or Cauchy–to be used to perturb free-parameter values. Defaults to the uniform distribution (for mxJiggle()).

loc, scale

Numeric. The location and scale parameters of the distribution from which random values are drawn to perturb free-parameter values, defaulting respectively to 1 and 0.25 (for mxJiggle()).

params

Numeric vector of current free parameter values.

lbounds

Numeric vector of lower bounds on parameters.

ubounds

Numeric vector of upper bounds on parameters.

Details

If mxJiggle() argument classic=FALSE (the default), mxJiggle() calls imxJiggle(). In that case, mxJiggle() passes imxJiggle() its own values for arguments dsn, loc, and scale, and extracts values for arguments params, lbounds, and ubounds from model. Then, model's free-parameter values are randomly perturbed before being re-assigned to it. The distributional family from which the perturbations are randomly generated is dictated by argument dsn. The distribution is parameterized by arguments loc and scale, respectively the location and scale parameters. The location parameter is the distribution's median. For the uniform distribution, scale is the absolute difference between its median and extrema (i.e., half the width of the rectangle); for the normal distribution, scale is its standard deviation; and for the Cauchy, scale is one-half its interquartile range. Free-parameter values are first multiplied by random draws from a distribution with the provided loc and scale, then added to random draws from a distribution with the same scale but with a median of zero.

If mxJiggle() argument classic=TRUE, then each free-parameter value x_i is replaced with x_i + 0.1(x_i + 0.5); this is the same behavior elicited in classic Mx by keyword JIGGLE.

Value

imxJiggle() returns a numeric vector of randomly perturbed free-parameter values. mxJiggle() returns model, with its free parameter values altered according to the other function arguments.

See Also

mxTryHard()

Examples

data(demoOneFactor)
manifests <- names(demoOneFactor)
latents <- c("G")
factorModel <- mxModel(
	"One Factor",
	type="RAM",
	manifestVars = manifests,
	latentVars = latents,
	mxPath(from=latents, to=manifests,values=0.8),
	mxPath(from=manifests, arrows=2,values=1),
	mxPath(from=latents, arrows=2,
				 free=FALSE, values=1.0),
	mxData(cov(demoOneFactor), type="cov",
				 numObs=500)
)

iniPars <- coef(factorModel)
print(iniPars)

pars2 <- imxJiggle(params=iniPars,lbounds=NA,ubounds=NA,dsn="runif",loc=1,scale=0.05)
print(pars2)

mod2 <- mxJiggle(model=factorModel,scale=0.05)
coef(mod2)

mod3 <- mxJiggle(model=factorModel,classic=TRUE)
coef(mod3)