Nothing
inst/models/nightly/ModificationIndexCheck.R
In OpenMx: Extended Structural Equation Modelling
#
# Copyright 2007-2019 by the individuals mentioned in the source code history
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#------------------------------------------------------------------------------
# Author: Michael D. Hunter
# Date: 2015-03-11
# Filename: ModificationIndexCheck.R
# Purpose: Check that modification indices are behaving reasonably.
#------------------------------------------------------------------------------
#------------------------------------------------------------------------------
# Factor model check
require(OpenMx)
vv <- c("No","Yes")
for(v in vv){
mxOption(NULL,"Analytic Gradients",v); mxOption(NULL,"Analytic RAM derivatives",v)
data(demoOneFactor)
manifests <- names(demoOneFactor)
latents <- c("G")
factorModel <- mxModel(
"One Factor",
type="RAM",
manifestVars = manifests,
latentVars = latents,
mxPath(from=latents, to=manifests),
mxPath(from=manifests, arrows=2),
mxPath(from=latents, arrows=2,
free=FALSE, values=1.0),
mxPath(from = 'one', to = manifests),
mxData(demoOneFactor, type="raw"))
factorRun <- mxRun(factorModel)
# This test script fails when CSOLNP and NPSOL are in use with analytic gradients. The reason is that they both reach a solution such that,
# when it comes time to compute the modification index for the sole fixed parameter in the 'M' matrix, the corresponding
# "full Hessian" is just barely uninvertible. This seems to simply be an edge case, due to "bad luck".
mm <- c("A","S","M")
if(mxOption(NULL,"Analytic gradients")=="Yes" && mxOption(NULL,"Default optimizer")!="SLSQP"){
mm <- c("A","S")
}
fim <- mxMI(factorRun,matrices=mm)
mplusON.mi <- c(NA, 0.001, 1.582, .272, .285, NA, 0.001, NA, 1.540, 4.041, .748, NA, 1.582, 1.540, NA, .151, 5.339, NA, .272, 4.042, .151, NA, 3.574, NA, NA, .748, 5.339, 3.574, NA, NA, NA)
mplusWITH.mi <- c(0.001, 1.582, 1.540, .272, NA, .151, .285, .748, 5.339, 3.574)
Adiag <- c(1L, 8L, 15L, 22L, 29L, 31L)
mplus.mi <- c(mplusON.mi[-Adiag], mplusWITH.mi)
plot(fim$MI.Full[1:length(mplus.mi)], mplus.mi)
abline(a=0, b=1)
points(fim$MI[1:length(mplus.mi)], mplus.mi, col='blue')
(theCor <- cor(fim$MI.Full[1:length(mplus.mi)], mplus.mi, use="pair"))
rms <- function(x, y){sqrt(mean((x-y)^2, na.rm=TRUE))}
(theRMS <- rms(fim$MI.Full[1:length(mplus.mi)], mplus.mi))
omxCheckTrue(theCor > 0.95)
omxCheckTrue(theRMS < 0.50)
omCheck <- fim$MI.Full[1:length(mplus.mi)][!is.na(mplus.mi)]
mpCheck <- mplus.mi[!is.na(mplus.mi)]
relMean <- mean(omCheck/mpCheck)
omxCheckTrue( relMean < 1.05 & relMean > 0.95)
#------------------------------------------------------------------------------
# State space model check
# Create data based on state space model.
require(OpenMx)
nvar <- 5
varnames <- paste("x", 1:nvar, sep="")
ssModel <- mxModel(
model="State Space Manual Example",
mxMatrix("Full", 1, 1, TRUE, .3, name="A"),
mxMatrix("Zero", 1, 1, name="B"),
mxMatrix("Full", nvar, 1, TRUE, .6, name="C", dimnames=list(varnames, "F1")),
mxMatrix("Zero", nvar, 1, name="D"),
mxMatrix("Diag", 1, 1, FALSE, 1, name="Q"),
mxMatrix("Diag", nvar, nvar, TRUE, .2, name="R"),
mxMatrix("Zero", 1, 1, name="x0"),
mxMatrix("Diag", 1, 1, FALSE, 1, name="P0"),
mxMatrix("Zero", 1, 1, name="u"),
mxExpectationStateSpace("A", "B", "C", "D", "Q", "R", "x0", "P0", "u"),
mxFitFunctionML()
)
Rmis <- ssModel$R
lsel <- lower.tri(Rmis$values, TRUE)
mval <- Rmis$values[lsel]
mval[2] <- .05
mfre <- Rmis$free[lsel]
mfre[2] <- TRUE
mlab <- Rmis$labels[lsel]
Rmis <- mxMatrix("Symm", nrow(Rmis), ncol(Rmis), values=mval, labels=mlab, free=mfre, name=Rmis$name)
ssMis <- mxModel(ssModel, Rmis)
set.seed(101)
ssMisData <- mxGenerateData(ssMis, 200)
ssMisRun <- mxRun(mxModel(ssModel, mxData(ssMisData, 'raw')))
mi.mis <- mxMI(ssMisRun, full=FALSE)
if( any(mi.mis$MI > qchisq(p=1-0.01, df=1), na.rm=TRUE)){
print("Large modification index found")
# Grab the plus.param model that has the highest modification index
newModelAttempt <- mi.mis$plusOneParamModels[[which.max(mi.mis$MI)]]
newRunAttempt <- mxRun(newModelAttempt)
mi.cor <- mxMI(newRunAttempt, full=FALSE)
foundFirst <- TRUE
if( any(mi.cor$MI > qchisq(p=1-0.01, df=1), na.rm=TRUE ) ){
print("Problem. Model still may need modification")
foundSecond <- TRUE
} else{
print("Success. All modification indices are sufficiently small now.")
foundSecond <- FALSE
}
fim <- mxMI(factorRun)
mplusON.mi <- c(NA, 0.001, 1.582, .272, .285, NA, 0.001, NA, 1.540, 4.041, .748, NA, 1.582, 1.540, NA, .151, 5.339, NA, .272, 4.042, .151, NA, 3.574, NA, NA, .748, 5.339, 3.574, NA, NA, NA)
mplusWITH.mi <- c(0.001, 1.582, 1.540, .272, NA, .151, .285, .748, 5.339, 3.574)
Adiag <- c(1L, 8L, 15L, 22L, 29L, 31L)
mplus.mi <- c(mplusON.mi[-Adiag], mplusWITH.mi)
plot(fim$MI.Full[1:length(mplus.mi)], mplus.mi)
abline(a=0, b=1)
points(fim$MI[1:length(mplus.mi)], mplus.mi, col='blue')
(theCor <- cor(fim$MI.Full[1:length(mplus.mi)], mplus.mi, use="pair"))
rms <- function(x, y){sqrt(mean((x-y)^2, na.rm=TRUE))}
(theRMS <- rms(fim$MI.Full[1:length(mplus.mi)], mplus.mi))
omxCheckTrue(theCor > 0.95)
omxCheckTrue(theRMS < 0.50)
omCheck <- fim$MI.Full[1:length(mplus.mi)][!is.na(mplus.mi)]
mpCheck <- mplus.mi[!is.na(mplus.mi)]
relMean <- mean(omCheck/mpCheck)
omxCheckTrue( relMean < 1.05 & relMean > 0.95)
#------------------------------------------------------------------------------
# State space model check
# Create data based on state space model.
require(OpenMx)
nvar <- 5
varnames <- paste("x", 1:nvar, sep="")
ssModel <- mxModel(model="State Space Manual Example",
mxMatrix("Full", 1, 1, TRUE, .3, name="A"),
mxMatrix("Zero", 1, 1, name="B"),
mxMatrix("Full", nvar, 1, TRUE, .6, name="C", dimnames=list(varnames, "F1")),
mxMatrix("Zero", nvar, 1, name="D"),
mxMatrix("Diag", 1, 1, FALSE, 1, name="Q"),
mxMatrix("Diag", nvar, nvar, TRUE, .2, name="R"),
mxMatrix("Zero", 1, 1, name="x0"),
mxMatrix("Diag", 1, 1, FALSE, 1, name="P0"),
mxMatrix("Zero", 1, 1, name="u"),
mxExpectationStateSpace("A", "B", "C", "D", "Q", "R", "x0", "P0", "u"),
mxFitFunctionML()
)
Rmis <- ssModel$R
lsel <- lower.tri(Rmis$values, TRUE)
mval <- Rmis$values[lsel]
mval[2] <- .05
mfre <- Rmis$free[lsel]
mfre[2] <- TRUE
mlab <- Rmis$labels[lsel]
Rmis <- mxMatrix("Symm", nrow(Rmis), ncol(Rmis), values=mval, labels=mlab, free=mfre, name=Rmis$name)
ssMis <- mxModel(ssModel, Rmis)
set.seed(101)
ssMisData <- mxGenerateData(ssMis, 400)
ssMisRun <- mxRun(mxModel(ssModel, mxData(ssMisData, 'raw')))
mi.mis <- mxMI(ssMisRun, full=FALSE)
if( any(mi.mis$MI > qchisq(p=0.05, df=1, lower.tail=FALSE), na.rm=TRUE)){
print("Large modification index found")
# Grab the plus.param model that has the highest modification index
newModelAttempt <- mi.mis$plusOneParamModels[[which.max(mi.mis$MI)]]
newRunAttempt <- mxRun(newModelAttempt)
mi.cor <- mxMI(newRunAttempt, full=FALSE)
foundFirst <- TRUE
if( any(mi.cor$MI > qchisq(p=0.05, df=1, lower.tail=FALSE), na.rm=TRUE ) ){
print("Problem. Model still may need modification")
foundSecond <- TRUE
} else{
foundFirst <- FALSE
}
omxCheckTrue(all( c(foundFirst, foundSecond) == c(TRUE, FALSE) ) )
prevMIsForSS <- c(0.000002, 10.827511, 1.664330, 2.062372, 1.211038,
1.428606, 3.021569, 1.749314, 0.752319, 2.663316, 1.167776,
0.032929, -0.186243)
miSSCheck <- mi.mis$MI[!is.na(mi.mis$MI)]
miSSCheck <- miSSCheck[-length(miSSCheck)]
prevMIsForSS <- prevMIsForSS[-length(prevMIsForSS)]
omxCheckCloseEnough(miSSCheck, prevMIsForSS, 0.01)
#------------------------------------------------------------------------------
# State space model check
# Similar to the above, but tests the respect of fixed parameter labels.
# When a fixed parameter occurs in more than one spot (and thus necessarily
# has a label), it is freed by its label instead of by the matrix element.
# Create data based on state space model.
require(OpenMx)
nvar <- 5
varnames <- paste("x", 1:nvar, sep="")
ssModel <- mxModel(
model="State Space Manual Example",
mxMatrix("Full", 1, 1, TRUE, .3, name="A"),
mxMatrix("Zero", 1, 1, name="B"),
mxMatrix("Full", nvar, 1, TRUE, .6, name="C", dimnames=list(varnames, "F1")),
mxMatrix("Zero", nvar, 1, name="D"),
mxMatrix("Diag", 1, 1, FALSE, 1, name="Q"),
mxMatrix("Diag", nvar, nvar, c(F, F, T, T, T), .2, name="R", labels=paste('resid', c(1, 1, 3:nvar), sep='')),
mxMatrix("Zero", 1, 1, name="x0"),
mxMatrix("Diag", 1, 1, FALSE, 1, name="P0"),
mxMatrix("Zero", 1, 1, name="u"),
mxExpectationStateSpace("A", "B", "C", "D", "Q", "R", "x0", "P0", "u"),
mxFitFunctionML()
)
Rmis <- ssModel$R
lsel <- lower.tri(Rmis$values, TRUE)
mval <- Rmis$values[lsel]
mval[2] <- .05
mfre <- Rmis$free[lsel]
mfre[2] <- TRUE
mlab <- Rmis$labels[lsel]
Rmis <- mxMatrix("Symm", nrow(Rmis), ncol(Rmis), values=mval, labels=mlab, free=mfre, name=Rmis$name)
ssMis <- mxModel(ssModel, Rmis)
set.seed(101)
ssMisData <- mxGenerateData(ssMis, 200)
ssMisRun <- mxRun(mxModel(ssModel, mxData(ssMisData, 'raw')))
mi.mis <- mxMI(ssMisRun, full=FALSE)
if( any(mi.mis$MI > qchisq(p=1-0.01, df=1), na.rm=TRUE)){
print("Large modification index found")
# Grab the plus.param model that has the highest modification index
newModelAttempt <- mi.mis$plusOneParamModels[[which.max(mi.mis$MI)]]
newRunAttempt <- mxRun(newModelAttempt)
mi.cor <- mxMI(newRunAttempt, full=FALSE)
foundFirst <- TRUE
if( any(mi.cor$MI > qchisq(p=1-0.01, df=1), na.rm=TRUE ) ){
print("Problem. Model still may need modification")
foundSecond <- TRUE
} else{
print("Success. All modification indices are sufficiently small now.")
foundSecond <- FALSE
}
} else{
foundFirst <- FALSE
}
omxCheckTrue(all( c(foundFirst, foundSecond) == c(TRUE, FALSE) ) )
prevMIsForSS <- c(0.000000, 5.201267, 1.857374, 0.681598, 1.678599,
0.247626, 0.962515, 0.018834, 4.511735, 3.853470, 0.000809, 0.032943)
miSSCheck <- mi.mis$MI[!is.na(mi.mis$MI)]
miSSCheck <- miSSCheck[-length(miSSCheck)]
omxCheckCloseEnough(miSSCheck, prevMIsForSS, 0.01)
#------------------------------------------------------------------------------
# State space model check
# Similar to the above, but tests the respect of fixed parameter labels.
# When a fixed parameter occurs in more than one spot (and thus necessarily
# has a label), it is freed by its label instead of by the matrix element.
# Create data based on state space model.
require(OpenMx)
nvar <- 5
varnames <- paste("x", 1:nvar, sep="")
ssModel <- mxModel(model="State Space Manual Example",
mxMatrix("Full", 1, 1, TRUE, .3, name="A"),
mxMatrix("Zero", 1, 1, name="B"),
mxMatrix("Full", nvar, 1, TRUE, .6, name="C", dimnames=list(varnames, "F1")),
mxMatrix("Zero", nvar, 1, name="D"),
mxMatrix("Diag", 1, 1, FALSE, 1, name="Q"),
mxMatrix("Diag", nvar, nvar, c(F, F, T, T, T), .2, name="R", labels=paste('resid', c(1, 1, 3:nvar), sep='')),
mxMatrix("Zero", 1, 1, name="x0"),
mxMatrix("Diag", 1, 1, FALSE, 1, name="P0"),
mxMatrix("Zero", 1, 1, name="u"),
mxExpectationStateSpace("A", "B", "C", "D", "Q", "R", "x0", "P0", "u"),
mxFitFunctionML()
)
Rmis <- ssModel$R
lsel <- lower.tri(Rmis$values, TRUE)
mval <- Rmis$values[lsel]
mval[2] <- .05
mfre <- Rmis$free[lsel]
mfre[2] <- TRUE
mlab <- Rmis$labels[lsel]
Rmis <- mxMatrix("Symm", nrow(Rmis), ncol(Rmis), values=mval, labels=mlab, free=mfre, name=Rmis$name)
ssMis <- mxModel(ssModel, Rmis)
set.seed(101)
ssMisData <- mxGenerateData(ssMis, 200)
ssMisRun <- mxRun(mxModel(ssModel, mxData(ssMisData, 'raw')))
mi.mis <- mxMI(ssMisRun, full=FALSE)
if( any(mi.mis$MI > qchisq(p=1-0.01, df=1), na.rm=TRUE)){
print("Large modification index found")
# Grab the plus.param model that has the highest modification index
newModelAttempt <- mi.mis$plusOneParamModels[[which.max(mi.mis$MI)]]
newRunAttempt <- mxRun(newModelAttempt)
mi.cor <- mxMI(newRunAttempt, full=FALSE)
foundFirst <- TRUE
if( any(mi.cor$MI > qchisq(p=1-0.01, df=1), na.rm=TRUE ) ){
print("Problem. Model still may need modification")
foundSecond <- TRUE
} else{
foundFirst <- FALSE
}
omxCheckTrue(all( c(foundFirst, foundSecond) == c(TRUE, FALSE) ) )
miSSCheck <- mi.mis$MI[!is.na(mi.mis$MI)]
miSSCheck <- miSSCheck['resid1']
prevMIsForSS <- 3.909195
omxCheckCloseEnough(miSSCheck, prevMIsForSS, 0.01)
}
}
}
}
mxOption(reset=TRUE)
Try the OpenMx package in your browser
Any scripts or data that you put into this service are public.
OpenMx documentation built on June 8, 2025, 9:33 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
#
# Copyright 2007-2019 by the individuals mentioned in the source code history
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#------------------------------------------------------------------------------
# Author: Michael D. Hunter
# Date: 2015-03-11
# Filename: ModificationIndexCheck.R
# Purpose: Check that modification indices are behaving reasonably.
#------------------------------------------------------------------------------
#------------------------------------------------------------------------------
# Factor model check
require(OpenMx)
vv <- c("No","Yes")
for(v in vv){
mxOption(NULL,"Analytic Gradients",v); mxOption(NULL,"Analytic RAM derivatives",v)
data(demoOneFactor)
manifests <- names(demoOneFactor)
latents <- c("G")
factorModel <- mxModel(
"One Factor",
type="RAM",
manifestVars = manifests,
latentVars = latents,
mxPath(from=latents, to=manifests),
mxPath(from=manifests, arrows=2),
mxPath(from=latents, arrows=2,
free=FALSE, values=1.0),
mxPath(from = 'one', to = manifests),
mxData(demoOneFactor, type="raw"))
factorRun <- mxRun(factorModel)
# This test script fails when CSOLNP and NPSOL are in use with analytic gradients. The reason is that they both reach a solution such that,
# when it comes time to compute the modification index for the sole fixed parameter in the 'M' matrix, the corresponding
# "full Hessian" is just barely uninvertible. This seems to simply be an edge case, due to "bad luck".
mm <- c("A","S","M")
if(mxOption(NULL,"Analytic gradients")=="Yes" && mxOption(NULL,"Default optimizer")!="SLSQP"){
mm <- c("A","S")
}
fim <- mxMI(factorRun,matrices=mm)
mplusON.mi <- c(NA, 0.001, 1.582, .272, .285, NA, 0.001, NA, 1.540, 4.041, .748, NA, 1.582, 1.540, NA, .151, 5.339, NA, .272, 4.042, .151, NA, 3.574, NA, NA, .748, 5.339, 3.574, NA, NA, NA)
mplusWITH.mi <- c(0.001, 1.582, 1.540, .272, NA, .151, .285, .748, 5.339, 3.574)
Adiag <- c(1L, 8L, 15L, 22L, 29L, 31L)
mplus.mi <- c(mplusON.mi[-Adiag], mplusWITH.mi)
plot(fim$MI.Full[1:length(mplus.mi)], mplus.mi)
abline(a=0, b=1)
points(fim$MI[1:length(mplus.mi)], mplus.mi, col='blue')
(theCor <- cor(fim$MI.Full[1:length(mplus.mi)], mplus.mi, use="pair"))
rms <- function(x, y){sqrt(mean((x-y)^2, na.rm=TRUE))}
(theRMS <- rms(fim$MI.Full[1:length(mplus.mi)], mplus.mi))
omxCheckTrue(theCor > 0.95)
omxCheckTrue(theRMS < 0.50)
omCheck <- fim$MI.Full[1:length(mplus.mi)][!is.na(mplus.mi)]
mpCheck <- mplus.mi[!is.na(mplus.mi)]
relMean <- mean(omCheck/mpCheck)
omxCheckTrue( relMean < 1.05 & relMean > 0.95)
#------------------------------------------------------------------------------
# State space model check
# Create data based on state space model.
require(OpenMx)
nvar <- 5
varnames <- paste("x", 1:nvar, sep="")
ssModel <- mxModel(
model="State Space Manual Example",
mxMatrix("Full", 1, 1, TRUE, .3, name="A"),
mxMatrix("Zero", 1, 1, name="B"),
mxMatrix("Full", nvar, 1, TRUE, .6, name="C", dimnames=list(varnames, "F1")),
mxMatrix("Zero", nvar, 1, name="D"),
mxMatrix("Diag", 1, 1, FALSE, 1, name="Q"),
mxMatrix("Diag", nvar, nvar, TRUE, .2, name="R"),
mxMatrix("Zero", 1, 1, name="x0"),
mxMatrix("Diag", 1, 1, FALSE, 1, name="P0"),
mxMatrix("Zero", 1, 1, name="u"),
mxExpectationStateSpace("A", "B", "C", "D", "Q", "R", "x0", "P0", "u"),
mxFitFunctionML()
)
Rmis <- ssModel$R
lsel <- lower.tri(Rmis$values, TRUE)
mval <- Rmis$values[lsel]
mval[2] <- .05
mfre <- Rmis$free[lsel]
mfre[2] <- TRUE
mlab <- Rmis$labels[lsel]
Rmis <- mxMatrix("Symm", nrow(Rmis), ncol(Rmis), values=mval, labels=mlab, free=mfre, name=Rmis$name)
ssMis <- mxModel(ssModel, Rmis)
set.seed(101)
ssMisData <- mxGenerateData(ssMis, 200)
ssMisRun <- mxRun(mxModel(ssModel, mxData(ssMisData, 'raw')))
mi.mis <- mxMI(ssMisRun, full=FALSE)
if( any(mi.mis$MI > qchisq(p=1-0.01, df=1), na.rm=TRUE)){
print("Large modification index found")
# Grab the plus.param model that has the highest modification index
newModelAttempt <- mi.mis$plusOneParamModels[[which.max(mi.mis$MI)]]
newRunAttempt <- mxRun(newModelAttempt)
mi.cor <- mxMI(newRunAttempt, full=FALSE)
foundFirst <- TRUE
if( any(mi.cor$MI > qchisq(p=1-0.01, df=1), na.rm=TRUE ) ){
print("Problem. Model still may need modification")
foundSecond <- TRUE
} else{
print("Success. All modification indices are sufficiently small now.")
foundSecond <- FALSE
}
fim <- mxMI(factorRun)
mplusON.mi <- c(NA, 0.001, 1.582, .272, .285, NA, 0.001, NA, 1.540, 4.041, .748, NA, 1.582, 1.540, NA, .151, 5.339, NA, .272, 4.042, .151, NA, 3.574, NA, NA, .748, 5.339, 3.574, NA, NA, NA)
mplusWITH.mi <- c(0.001, 1.582, 1.540, .272, NA, .151, .285, .748, 5.339, 3.574)
Adiag <- c(1L, 8L, 15L, 22L, 29L, 31L)
mplus.mi <- c(mplusON.mi[-Adiag], mplusWITH.mi)
plot(fim$MI.Full[1:length(mplus.mi)], mplus.mi)
abline(a=0, b=1)
points(fim$MI[1:length(mplus.mi)], mplus.mi, col='blue')
(theCor <- cor(fim$MI.Full[1:length(mplus.mi)], mplus.mi, use="pair"))
rms <- function(x, y){sqrt(mean((x-y)^2, na.rm=TRUE))}
(theRMS <- rms(fim$MI.Full[1:length(mplus.mi)], mplus.mi))
omxCheckTrue(theCor > 0.95)
omxCheckTrue(theRMS < 0.50)
omCheck <- fim$MI.Full[1:length(mplus.mi)][!is.na(mplus.mi)]
mpCheck <- mplus.mi[!is.na(mplus.mi)]
relMean <- mean(omCheck/mpCheck)
omxCheckTrue( relMean < 1.05 & relMean > 0.95)
#------------------------------------------------------------------------------
# State space model check
# Create data based on state space model.
require(OpenMx)
nvar <- 5
varnames <- paste("x", 1:nvar, sep="")
ssModel <- mxModel(model="State Space Manual Example",
mxMatrix("Full", 1, 1, TRUE, .3, name="A"),
mxMatrix("Zero", 1, 1, name="B"),
mxMatrix("Full", nvar, 1, TRUE, .6, name="C", dimnames=list(varnames, "F1")),
mxMatrix("Zero", nvar, 1, name="D"),
mxMatrix("Diag", 1, 1, FALSE, 1, name="Q"),
mxMatrix("Diag", nvar, nvar, TRUE, .2, name="R"),
mxMatrix("Zero", 1, 1, name="x0"),
mxMatrix("Diag", 1, 1, FALSE, 1, name="P0"),
mxMatrix("Zero", 1, 1, name="u"),
mxExpectationStateSpace("A", "B", "C", "D", "Q", "R", "x0", "P0", "u"),
mxFitFunctionML()
)
Rmis <- ssModel$R
lsel <- lower.tri(Rmis$values, TRUE)
mval <- Rmis$values[lsel]
mval[2] <- .05
mfre <- Rmis$free[lsel]
mfre[2] <- TRUE
mlab <- Rmis$labels[lsel]
Rmis <- mxMatrix("Symm", nrow(Rmis), ncol(Rmis), values=mval, labels=mlab, free=mfre, name=Rmis$name)
ssMis <- mxModel(ssModel, Rmis)
set.seed(101)
ssMisData <- mxGenerateData(ssMis, 400)
ssMisRun <- mxRun(mxModel(ssModel, mxData(ssMisData, 'raw')))
mi.mis <- mxMI(ssMisRun, full=FALSE)
if( any(mi.mis$MI > qchisq(p=0.05, df=1, lower.tail=FALSE), na.rm=TRUE)){
print("Large modification index found")
# Grab the plus.param model that has the highest modification index
newModelAttempt <- mi.mis$plusOneParamModels[[which.max(mi.mis$MI)]]
newRunAttempt <- mxRun(newModelAttempt)
mi.cor <- mxMI(newRunAttempt, full=FALSE)
foundFirst <- TRUE
if( any(mi.cor$MI > qchisq(p=0.05, df=1, lower.tail=FALSE), na.rm=TRUE ) ){
print("Problem. Model still may need modification")
foundSecond <- TRUE
} else{
foundFirst <- FALSE
}
omxCheckTrue(all( c(foundFirst, foundSecond) == c(TRUE, FALSE) ) )
prevMIsForSS <- c(0.000002, 10.827511, 1.664330, 2.062372, 1.211038,
1.428606, 3.021569, 1.749314, 0.752319, 2.663316, 1.167776,
0.032929, -0.186243)
miSSCheck <- mi.mis$MI[!is.na(mi.mis$MI)]
miSSCheck <- miSSCheck[-length(miSSCheck)]
prevMIsForSS <- prevMIsForSS[-length(prevMIsForSS)]
omxCheckCloseEnough(miSSCheck, prevMIsForSS, 0.01)
#------------------------------------------------------------------------------
# State space model check
# Similar to the above, but tests the respect of fixed parameter labels.
# When a fixed parameter occurs in more than one spot (and thus necessarily
# has a label), it is freed by its label instead of by the matrix element.
# Create data based on state space model.
require(OpenMx)
nvar <- 5
varnames <- paste("x", 1:nvar, sep="")
ssModel <- mxModel(
model="State Space Manual Example",
mxMatrix("Full", 1, 1, TRUE, .3, name="A"),
mxMatrix("Zero", 1, 1, name="B"),
mxMatrix("Full", nvar, 1, TRUE, .6, name="C", dimnames=list(varnames, "F1")),
mxMatrix("Zero", nvar, 1, name="D"),
mxMatrix("Diag", 1, 1, FALSE, 1, name="Q"),
mxMatrix("Diag", nvar, nvar, c(F, F, T, T, T), .2, name="R", labels=paste('resid', c(1, 1, 3:nvar), sep='')),
mxMatrix("Zero", 1, 1, name="x0"),
mxMatrix("Diag", 1, 1, FALSE, 1, name="P0"),
mxMatrix("Zero", 1, 1, name="u"),
mxExpectationStateSpace("A", "B", "C", "D", "Q", "R", "x0", "P0", "u"),
mxFitFunctionML()
)
Rmis <- ssModel$R
lsel <- lower.tri(Rmis$values, TRUE)
mval <- Rmis$values[lsel]
mval[2] <- .05
mfre <- Rmis$free[lsel]
mfre[2] <- TRUE
mlab <- Rmis$labels[lsel]
Rmis <- mxMatrix("Symm", nrow(Rmis), ncol(Rmis), values=mval, labels=mlab, free=mfre, name=Rmis$name)
ssMis <- mxModel(ssModel, Rmis)
set.seed(101)
ssMisData <- mxGenerateData(ssMis, 200)
ssMisRun <- mxRun(mxModel(ssModel, mxData(ssMisData, 'raw')))
mi.mis <- mxMI(ssMisRun, full=FALSE)
if( any(mi.mis$MI > qchisq(p=1-0.01, df=1), na.rm=TRUE)){
print("Large modification index found")
# Grab the plus.param model that has the highest modification index
newModelAttempt <- mi.mis$plusOneParamModels[[which.max(mi.mis$MI)]]
newRunAttempt <- mxRun(newModelAttempt)
mi.cor <- mxMI(newRunAttempt, full=FALSE)
foundFirst <- TRUE
if( any(mi.cor$MI > qchisq(p=1-0.01, df=1), na.rm=TRUE ) ){
print("Problem. Model still may need modification")
foundSecond <- TRUE
} else{
print("Success. All modification indices are sufficiently small now.")
foundSecond <- FALSE
}
} else{
foundFirst <- FALSE
}
omxCheckTrue(all( c(foundFirst, foundSecond) == c(TRUE, FALSE) ) )
prevMIsForSS <- c(0.000000, 5.201267, 1.857374, 0.681598, 1.678599,
0.247626, 0.962515, 0.018834, 4.511735, 3.853470, 0.000809, 0.032943)
miSSCheck <- mi.mis$MI[!is.na(mi.mis$MI)]
miSSCheck <- miSSCheck[-length(miSSCheck)]
omxCheckCloseEnough(miSSCheck, prevMIsForSS, 0.01)
#------------------------------------------------------------------------------
# State space model check
# Similar to the above, but tests the respect of fixed parameter labels.
# When a fixed parameter occurs in more than one spot (and thus necessarily
# has a label), it is freed by its label instead of by the matrix element.
# Create data based on state space model.
require(OpenMx)
nvar <- 5
varnames <- paste("x", 1:nvar, sep="")
ssModel <- mxModel(model="State Space Manual Example",
mxMatrix("Full", 1, 1, TRUE, .3, name="A"),
mxMatrix("Zero", 1, 1, name="B"),
mxMatrix("Full", nvar, 1, TRUE, .6, name="C", dimnames=list(varnames, "F1")),
mxMatrix("Zero", nvar, 1, name="D"),
mxMatrix("Diag", 1, 1, FALSE, 1, name="Q"),
mxMatrix("Diag", nvar, nvar, c(F, F, T, T, T), .2, name="R", labels=paste('resid', c(1, 1, 3:nvar), sep='')),
mxMatrix("Zero", 1, 1, name="x0"),
mxMatrix("Diag", 1, 1, FALSE, 1, name="P0"),
mxMatrix("Zero", 1, 1, name="u"),
mxExpectationStateSpace("A", "B", "C", "D", "Q", "R", "x0", "P0", "u"),
mxFitFunctionML()
)
Rmis <- ssModel$R
lsel <- lower.tri(Rmis$values, TRUE)
mval <- Rmis$values[lsel]
mval[2] <- .05
mfre <- Rmis$free[lsel]
mfre[2] <- TRUE
mlab <- Rmis$labels[lsel]
Rmis <- mxMatrix("Symm", nrow(Rmis), ncol(Rmis), values=mval, labels=mlab, free=mfre, name=Rmis$name)
ssMis <- mxModel(ssModel, Rmis)
set.seed(101)
ssMisData <- mxGenerateData(ssMis, 200)
ssMisRun <- mxRun(mxModel(ssModel, mxData(ssMisData, 'raw')))
mi.mis <- mxMI(ssMisRun, full=FALSE)
if( any(mi.mis$MI > qchisq(p=1-0.01, df=1), na.rm=TRUE)){
print("Large modification index found")
# Grab the plus.param model that has the highest modification index
newModelAttempt <- mi.mis$plusOneParamModels[[which.max(mi.mis$MI)]]
newRunAttempt <- mxRun(newModelAttempt)
mi.cor <- mxMI(newRunAttempt, full=FALSE)
foundFirst <- TRUE
if( any(mi.cor$MI > qchisq(p=1-0.01, df=1), na.rm=TRUE ) ){
print("Problem. Model still may need modification")
foundSecond <- TRUE
} else{
foundFirst <- FALSE
}
omxCheckTrue(all( c(foundFirst, foundSecond) == c(TRUE, FALSE) ) )
miSSCheck <- mi.mis$MI[!is.na(mi.mis$MI)]
miSSCheck <- miSSCheck['resid1']
prevMIsForSS <- 3.909195
omxCheckCloseEnough(miSSCheck, prevMIsForSS, 0.01)
}
}
}
}
mxOption(reset=TRUE)
Try the OpenMx package in your browser
Any scripts or data that you put into this service are public.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.