lava (version 1.5)

gof: Extract model summaries and GOF statistics for model object

Description

Calculates various GOF statistics for model object including global chi-squared test statistic and AIC. Extract model-specific mean and variance structure, residuals and various predicitions.

Usage

gof(object, ...)
# S3 method for lvmfit
gof(object, chisq=FALSE, level=0.90, rmsea.threshold=0.05,all=FALSE,...)
moments(x,...)
# S3 method for lvm
moments(x, p, debug=FALSE, conditional=FALSE, data=NULL, latent=FALSE, ...)
# S3 method for lvmfit
logLik(object, p=coef(object),
                      data=model.frame(object),
                      model=object$estimator,
                      weights=Weights(object),
                      data2=object$data$data2,
                          ...)
# S3 method for lvmfit
score(x, data=model.frame(x), p=pars(x), model=x$estimator,
                   weights=Weights(x), data2=x$data$data2, ...)
# S3 method for lvmfit
information(x,p=pars(x),n=x$data$n,data=model.frame(x),
                   model=x$estimator,weights=Weights(x), data2=x$data$data2, ...)

Arguments

object

Model object

…

Additional arguments to be passed to the low level functions

x

Model object

p

Parameter vector used to calculate statistics

data

Data.frame to use

latent

If TRUE predictions of latent variables are included in output

data2

Optional second data.frame (only for censored observations)

weights

Optional weight matrix

n

Number of observations

conditional

If TRUE the conditional moments given the covariates are calculated. Otherwise the joint moments are calculated

model

String defining estimator, e.g. "gaussian" (see estimate)

debug

Debugging only

chisq

Boolean indicating whether to calculate chi-squared goodness-of-fit (always TRUE for estimator='gaussian')

level

Level of confidence limits for RMSEA

rmsea.threshold

Which probability to calculate, Pr(RMSEA<rmsea.treshold)

all

Calculate all (ad hoc) FIT indices: TLI, CFI, NFI, SRMR, ...

Value

A htest-object.

Examples

m <- lvm(list(y~v1+v2+v3+v4,c(v1,v2,v3,v4)~x))
set.seed(1)
dd <- sim(m,1000)
e <- estimate(m, dd)
gof(e,all=TRUE,rmsea.threshold=0.05,level=0.9)


set.seed(1)
m <- lvm(list(c(y1,y2,y3)~u,y1~x)); latent(m) <- ~u
regression(m,c(y2,y3)~u) <- "b"
d <- sim(m,1000)
e <- estimate(m,d)
rsq(e)
##'
rr <- rsq(e,TRUE)
rr
estimate(rr,contrast=rbind(c(1,-1,0),c(1,0,-1),c(0,1,-1)))