SCM

Forum: help

Monitor Forum | Start New Thread Start New Thread
RE: nlssystemfit using SUR method [ reply ]
By: Arne Henningsen on 2014-09-14 06:59
[forum:41452]
Dear Francisco

I think that nlsystemfit() can estimate your model only if you assume a linear probability model for the survival probability model. Please also read the FAQ at systemfit's website:

http://systemfit.org/

I think that you need to derive the log-likelihood function (and preferably also its first derivatives) and maximise this function, e.g. with packages/function maxLik.

Best regards,
Arne

nlssystemfit using SUR method [ reply ]
By: Francisco Goes on 2014-08-31 10:39
[forum:41408]
For my master thesis I have fitted an individual tree diameter growth model and a survival probability model separately using R, but I was told that simultaneous estimation of these two models would minimize overall errors and provide a variance-covariance matrix as a whole.

In that respect, can you please tell me if I can do it with nlssystemfit using SUR (seemingly unrelated regression) method? If not, how can I do it in R?

My equations are:

- tree diameter growth model

d_richards_k<-d2~A*(1-exp(-(k0+k1*puro+k2*GL1/100+k3*1/G1+k4*Tmedmax/100+k5*Perc_G_ec1))*(1-(d1/A)^(1-m)))^(1/(1-m))
nls_d_richards_k<-nlsLM(d_richards_k,start=list(A=100,k0=0.6,k1=0,k2=0,k3=0,k4=0,k5=0,m=0.6),control=nls.lm.control(maxiter=500))

- survival probability model mortal_final<-glm(sobrev~Verao+alto_fuste:puro+IC1arv+G1+Perc_G_ec1,family=binomial)

Thank you!

Thanks to:
Vienna University of Economics and Business University of Wisconsin - Madison Powered By FusionForge