OpenMx: mxComputeGradientDescent – R documentation

Become an expert in R — Interactive courses, Cheat Sheets, certificates and more!

mxComputeGradientDescent

Optimize parameters using a gradient descent optimizer

Description

This optimizer does not require analytic derivatives of the fit function. The fully open-source CRAN version of OpenMx offers 2 choices, CSOLNP and SLSQP (from the NLOPT collection). The OpenMx Team's version of OpenMx offers the choice of three optimizers: CSOLNP, SLSQP, and NPSOL.

Usage

mxComputeGradientDescent(
  freeSet = NA_character_,
  ...,
  engine = NULL,
  fitfunction = "fitfunction",
  verbose = 0L,
  tolerance = NA_real_,
  useGradient = deprecated(),
  warmStart = NULL,
  nudgeZeroStarts = mxOption(NULL, "Nudge zero starts"),
  maxMajorIter = NULL,
  gradientAlgo = deprecated(),
  gradientIterations = deprecated(),
  gradientStepSize = deprecated()
)

Arguments

`freeSet`	names of matrices containing free parameters.
`...`	Not used. Forces remaining arguments to be specified by name.
`engine`	specific 'CSOLNP', 'SLSQP', or 'NPSOL'
`fitfunction`	name of the fitfunction (defaults to 'fitfunction')
`verbose`	integer. Level of run-time diagnostic output. Set to zero to disable
`tolerance`	how close to the optimum is close enough (also known as the optimality tolerance)
`useGradient`
`warmStart`	a Cholesky factored Hessian to use as the NPSOL Hessian starting value (preconditioner)
`nudgeZeroStarts`	whether to nudge any zero starting values prior to optimization (default TRUE)
`maxMajorIter`	maximum number of major iterations
`gradientAlgo`
`gradientIterations`
`gradientStepSize`

Details

All three optimizers can use analytic gradients, and only NPSOL uses warmStart. To customize more options, see mxOption.

References

Luenberger, D. G. & Ye, Y. (2008). Linear and nonlinear programming. Springer.

Examples

data(demoOneFactor)
factorModel <- mxModel(name ="One Factor",
  mxMatrix(type="Full", nrow=5, ncol=1, free=FALSE, values=0.2, name="A"),
    mxMatrix(type="Symm", nrow=1, ncol=1, free=FALSE, values=1, name="L"),
    mxMatrix(type="Diag", nrow=5, ncol=5, free=TRUE, values=1, name="U"),
    mxAlgebra(expression=A %*% L %*% t(A) + U, name="R"),
  mxExpectationNormal(covariance="R", dimnames=names(demoOneFactor)),
  mxFitFunctionML(),
    mxData(observed=cov(demoOneFactor), type="cov", numObs=500),
     mxComputeSequence(steps=list(
     mxComputeGradientDescent(),
     mxComputeNumericDeriv(),
     mxComputeStandardError(),
     mxComputeHessianQuality()
    )))
factorModelFit <- mxRun(factorModel)
factorModelFit$output$conditionNumber # 29.5

OpenMx

Extended Structural Equation Modelling

v2.19.5

Apache License (== 2.0)

Authors

Steven M. Boker [aut], Michael C. Neale [aut], Hermine H. Maes [aut], Michael J. Wilde [ctb], Michael Spiegel [aut], Timothy R. Brick [aut], Ryne Estabrook [aut], Timothy C. Bates [aut], Paras Mehta [ctb], Timo von Oertzen [ctb], Ross J. Gore [aut], Michael D. Hunter [aut], Daniel C. Hackett [ctb], Julian Karch [ctb], Andreas M. Brandmaier [ctb], Joshua N. Pritikin [aut, cre], Mahsa Zahery [aut], Robert M. Kirkpatrick [aut], Yang Wang [ctb], Ben Goodrich [ctb], Charles Driver [ctb], Massachusetts Institute of Technology [cph], S. G. Johnson [cph], Association for Computing Machinery [cph], Dieter Kraft [cph], Stefan Wilhelm [cph], Sarah Medland [cph], Carl F. Falk [cph], Matt Keller [cph], Manjunath B G [cph], The Regents of the University of California [cph], Lester Ingber [cph], Wong Shao Voon [cph], Juan Palacios [cph], Jiang Yang [cph], Gael Guennebaud [cph], Jitse Niesen [cph]

Initial release

2021-03-26