Displaying 1 result from an estimated 1 matches for "slopeabove".
2006 Aug 02
0
Trying to use segmented in a function
...iving me mad. Here's a simple example:
simdata<-function
(Ns=200,Xmean=20,Xsd=5,SdYerr=0.5,Yint=0,threshold=20,slopebelow=0.5,slo
peabove=1)
{
Xs<-rnorm(Ns,Xmean,Xsd)
Yerr<-rnorm(Ns,0,SdYerr)
D<-ifelse(Xs<=threshold,0,1)
XminusX0<-Xs-threshold
Ys<-Yint+slopebelow*Xs+slopeabove*XminusX0*D+Yerr
plot(Xs,Ys)
linmod<-lm(Ys~Xs)
segment<-segmented(linmod,Z=Xs,psi=threshold)
segment
}
This code should simply simulate some "breakpoint" data, with the
change in slope at "threshold" and then fit a model with segmented.
If I just use the code...