regMultiGroup: regMultiGroup

In jhorzek/regmx:

regMultiGroup

Description

Note: regmx is based on the R package regsem. Because of the early status of regmx, it is recommended to use regsem instead! regMultiGroup a mutliple Group Model with a penalty on the difference in parameter estimates between the groups. Its is not tested and should not be used!

Usage

regMultiGroup(
  mxModelObjects,
  regOn,
  regType = "lasso",
  regIndicators,
  regValue = 0
)

Arguments

mxModelObjects	a list with mxModels for the different groups
regOn	string vector with matrices that should be regularized. The matrices must have the same name as the ones provided in the mxModelObjects (e.g., "A").
regType	so far only "lasso" implemented
regIndicators	list of matrices indicating which parameters to regularize in the matrices provided in regOn. The matrices in regIndicators must to have the same names as the matrices they correspond to (e.g., regIndicators = list("A" = diag(10))). 1 Indicates a parameter that will be regularized, 0 an unregularized parameter
regValue	numeric value depicting the penalty size

Author(s)

Jannik Orzek

Examples

# The following example is adapted from the OpenMx documentation
library(OpenMx)

# Simulate some data

# Group 1
N1 = 100
x = rnorm(N1, mean= 0, sd= 1)
y = 0.5*x + rnorm(N1, mean= 0, sd= 1)
ds1 <- data.frame(x, y)
dsNames <- names(ds1)

# Group 2: higher regression weight
N2= 1000
x= rnorm(N2, mean= 0, sd= 1)
y= 0.6*x + rnorm(N2, mean= 0, sd= 1)
ds2 <- data.frame(x, y)


# Define the matrices (A matrix implementation of 2 RAM models)

# I is identical for both groups:
I <- mxMatrix(name="I", type="Iden", nrow=2, ncol=2)
# means are identical for both groups:
M <- mxMatrix(name = "M", type = "Full", nrow = 1, ncol = 2, values=0,
              free=TRUE, labels=c("Mean_x", "Mean_y"))
# The A matrix can differ:
A1 <- mxMatrix(name = "A", type = "Full", nrow = 2, ncol = 2, values=c(0,1,0,0),
               free=c(FALSE,TRUE,FALSE,FALSE), labels=c(NA, "b1", NA, NA))
A2 <- mxMatrix(name = "A", type = "Full", nrow = 2, ncol = 2, values=c(0,1,0,0),
               free=c(FALSE,TRUE,FALSE,FALSE), labels=c(NA, "b2", NA, NA))
# S is identical:
S <- mxMatrix(name = "S", type = "Diag", nrow = 2, ncol = 2, values=1,
              free=TRUE, labels=c("Var_x", "Resid"))


# Define the expectation
expect <- mxExpectationRAM('A', 'S', 'I', 'M', dimnames= dsNames)

# Choose a fit function
fitFunction <- mxFitFunctionML(rowDiagnostics=TRUE)
# Also return row likelihoods (the fit function value is still 1x1)

# Multiple-group fit function sums the model likelihoods
# from its component models
mgFitFun <- mxFitFunctionMultigroup(c('g1model', 'g2model'))


# Define model 1 and model 2
m1 = mxModel(model="g1model",
             M, S, A1, I, expect, fitFunction,
             mxData(cov(ds1), type="cov", numObs=N1, means=colMeans(ds1))
)
m2 = mxModel(model="g2model",
             M, S, A2, I, expect, fitFunction,
             mxData(cov(ds2), type="cov", numObs=N2, means=colMeans(ds2))
)

mg <- mxModel(model='multipleGroup', m1, m2, mgFitFun)
# All paths except for b are constrained to equality

# Fit the model and print a summary
mg <- mxRun(mg)
summary(mg)

############### Regularize A matrix:

regIndicators <- list("A" = matrix(c(0,1,1,0), nrow = 2))

regValue <- 1

regOn <- "A"

mxModelObjects <- list("model1" = m1, "model2" = m2)
regModel <- regMultiGroup(mxModelObjects = mxModelObjects, regOn = regOn, regIndicators = regIndicators, regType = "lasso",regValue = regValue )
fit_regModel <- mxRun(regModel)
summary(fit_regModel)

jhorzek/regmx documentation built on Sept. 19, 2022, 2:30 a.m.