ich nutze das Package VCA.
Dort gibt es eine Beispielberechnung mit einem Beispiel-Datensatz
Code: Alles auswählen
install.packages("VCA")
library("VCA")
# Function converts a color-string into RGB-code
# col (character) string specifying an R-color
# alpha (numeric) degree of transparency in [0, 1], 0=fully transparency, 1=opaque
asRGB <- function(col, alpha)
rgb(t(col2rgb(col))/255, alpha=alpha)
data(dataEP05A2_1)
varPlot(y~day/run, dataEP05A2_1,
# controls horizontal mean lines
MeanLine=list(var=c("int", "day"), col=c("gray75", "blue"), lwd=c(2,2)),
# controls how points (concentrations) are plotted, here using semi-transparency
# to see overlayed points
Points=list(pch=16, col=asRGB("black", .5), cex=1.25),
# controls how replicate-means are plotted
Mean=list(col="magenta", cex=1.25, lwd=2),
# controls how the title is shown
Title=list(main="20 x 2 x 2 Single-Site Evaluation", cex.main=1.75),
# controls plotting of levels per VC, if as many lists as there are VCs are
# specified, each VC can be specified individually
VarLab=list(list(cex=1.5), list(cex=1.25)),
# controls how names of VCs are plotted
VCnam=list(font=2, cex=1.5),
# controls appearance of the Y-axis label
YLabel=list(text="Concentation [mg/dL]", las=0, line=3, font=2, cex=1.25),
# Y-axis labels rotated
las=1)
# fit 20 x 2 x 2 model to data
fit.SS1 <- fitVCA(y~day/run, dataEP05A2_1)
fit.SS1
# estimate 95% confidence intervals, request CI for
# all variance components via 'VarVC=TRUE'
inf.SS1 <- VCAinference(fit.SS1, VarVC=TRUE)
inf.SS1
summary(inf.SS1)[1][[1]][[1]][[1]][[1]]
Das scheint aber nicht so einfach zu sein wie bei einem normalen linearen Modell.
Wir kann ich einfach herausfinden wo die Werte der "summary" also z.B. CI LCL, Estimate usw. liegen?
Vielen Dank für eure Hilfe
VG
Andreas