Hi,
Im trying to estimate different parameters in an alternative wrapped Cauchy
distribution (where mu (mui) is dependent on several variables (such as c, z
and mu0), where c is a constant) and using the following log-likelihood and
MLE functions:
###log likelihood for wrapped cauchy ###
wrappedCauchy.lik<-function(theta,y,x,z){
n<-nrow(y)
mu0 <- theta[1]
rho <- theta[2]
c <- theta[3]
mui <- mu0+(x-mu0)*(1-exp(-c*z))
logl<-sum(log((1-(rho^2))/(1+(rho^2)-2*rho*cos(y-mui))/(2*pi)))
return(-logl)
}
##Method for MLE###
d <- optim(c(mu0,rho0, 2.3),wrappedCauchy.lik,y=turn1,x = HRturns1, z =
HRDistance1, method="CG")
The problem here is that the optim results are equal to the initial values
and therefore changes dependent on those values am I doing something wrong
here?
When I run a simpler version of the wrapped cauchy just estimating mu and
rho (without mu being dependent on c), the optim estimates seem to work as
the are different from the initial values. Is it a problem trying to get MLE
for a constant in this case c?
Thanks
Lene
___________________________
Lene Jung Kjær
Studsdalvej 20, Taulov
7000 Fredericia, Danmark
Tlf: 29 86 96 14
email: ljk...@hotmail.com
[[alternative HTML version deleted]]
_______________________________________________
R-sig-ecology mailing list
R-sig-ecology@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-ecology
