lava (version 1.6.4)

twostageCV: Cross-validated two-stage estimator

Description

Cross-validated two-stage estimator for non-linear SEM

Usage

twostageCV(model1, model2, data, control1 = list(trace = 0),
  control2 = list(trace = 0), knots.boundary, mc.cores = 1,
  nmix = 1:4, df = 1:9, fix = TRUE, std.err = TRUE, nfolds = 5,
  rep = 1, messages = 0, ...)

Arguments

model1

model 1 (exposure measurement error model)

model2

model 2

data

data.frame

control1

optimization parameters for model 1

control2

optimization parameters for model 1

knots.boundary

boundary points for natural cubic spline basis

mc.cores

number of cores to use for parallel computations

nmix

number of mixture components

df

spline degrees of freedom

fix

automatically fix parameters for identification (TRUE)

std.err

calculation of standard errors (TRUE)

nfolds

Number of folds (cross-validation)

rep

Number of repeats of cross-validation

messages

print information (>0)

...

additional arguments to lower level functions

Examples

# NOT RUN {
 ## Reduce Ex.Timings
m1 <- lvm( x1+x2+x3 ~ u1, latent= ~u1)
m2 <- lvm( y ~ 1 )
m <- functional(merge(m1,m2), y ~ u, value=function(x) sin(x)+x)
distribution(m, ~u1) <- uniform.lvm(-6,6)
d <- sim(m,n=500,seed=1)
nonlinear(m2) <- y~u1
val <- twostageCV(m1, m2, data=d, std.err=FALSE, df=2:6, nmix=1:2,
                  nfolds=2, mc.cores=1)
val
# }