createARCLModel: createARCLModel

In jhorzek/OrzekMasterThesis: OrzekMasterThesis can be used to replicate the findings reportet in the master thesis "LASSO Regularization in Dynamic Panel Models"

Description

createARCLModel creates an ARCL(1)-SEM model with OpenMx with >= 1 latent variables and 1 manifest per latent and time point for data simulation. It returns the covariance based and a raw data based model

Usage

createARCLModel(sampleSize, numLatent, Timepoints, burning, Avalues, Afree,
  Alabel, Svalues, Sfree, Slabel, S_firstObsAllFree = TRUE,
  SimulatedDataSet)

Arguments

sampleSize	sample size
numLatent	number of latent varaibles per time point
Timepoints	number of time points observed in the final data set
burning	number of observations to burn before the initial time point
Avalues	values in ARCL matrix
Afree	free parameters in ARCL matrix
Alabel	labels parameters for ARCL matrix
Svalues	values in S matrix
Sfree	free parameters in S matrix
Slabel	labels parameters for S matrix
S_firstObsAllFree	should all initial observations be allowed to correlate?

Author(s)

Jannik Orzek

Examples

library(OpenMx)

Alabel <- matrix(c("a11", "a12", "a13", "a14", "a15", "a21", "a22", "a23", "a24", "a25", "a31", "a32", "a33", "a34", "a35", "a41", "a42", "a43", "a44", "a45", "a51", "a52", "a53", "a54", "a55"), nrow = 5, byrow = T)

Avalues <- diag(.5,nrow=5,ncol = 5)
Avalues[2:5,1] <-.3
Avalues[3:5,2] <-.3
Avalues[4,5] <-.3
Afree <- matrix(TRUE, 5,5)
Alabel <- matrix(c("a11", "a12", "a13", "a14", "a15", "a21", "a22", "a23", "a24", "a25", "a31", "a32", "a33", "a34", "a35", "a41", "a42", "a43", "a44", "a45", "a51", "a52", "a53", "a54", "a55"), nrow = 5, byrow = T)

Svalues <- diag(.75,nrow=5,ncol = 5)
Sfree <- diag(TRUE, 5,5)
Slabel <- matrix(c("s11", NA, NA, NA, NA,
                   NA, "s22", NA, NA, NA,
                   NA, NA, "s33", NA, NA,
                   NA, NA, NA, "s44", NA,
                   NA, NA, NA, NA, "s55"), nrow = 5, byrow = T)


Svalues_init <- diag(1,nrow=5,ncol = 5)
Sfree_init <- diag(TRUE, 5,5)
Slabel_init <- matrix(NA, nrow = 5, byrow = T)

simObj <- simARCL(numLatent = 5, Timepoints = 4, burning = 10,
                  Avalues = Avalues, Alabel = Alabel, Afree = Afree,
                  Svalues = Svalues, Sfree = Sfree, Slabel = Slabel,
                  Svalues_init = Svalues_init, Sfree_init = Sfree_init, Slabel_init = Slabel_init,
                  sampleSize = 1000, S_firstObsAllFree = T)


temp <- mxRun(simObj$mxARCL_cov)
summary(temp)

temp2 <- mxTryHard(simObj$mxARCL_FIML)
summary(temp2)

jhorzek/OrzekMasterThesis documentation built on Aug. 7, 2019, 8:20 a.m.