[R] need help in using Hessian matrix

Uwe Ligges ligges at statistik.tu-dortmund.de
Tue Nov 3 18:45:12 CET 2009


Laila Alkhalfan wrote:
> Hi
> I need to find the Hessian matrix for a complicated function from a certain
> kind of data but i keep getting this error
> Error in f1 - f2 : non-numeric argument to binary operator
> 
> the data is given by
> 
>           U<-runif(n)
>           Us<-sort(U)
>           tau1<- 2
>           F1tau<- pgamma((tau1/theta1),shape,1)
>           N1<-sum(Us<F1tau)
>           X1<- Us[1:N1]
>           Ti<- theta1*qgamma(Us[1:N1],shape,1)
>           w=N1+1
>           V<- Us[w:r]
>           N2<- length(V)
>           V1<-theta2*qgamma(V,shape,1)
>           Tj<-V1+tau1-(theta2/theta1)*tau1
>           c1<-matrix(Ti,ncol=1)
>           c2<-matrix(Tj,ncol=1)
>           cc<-data.frame(rbind(c1,c2))[,1]
> 
> require(numDeriv)
> m1<- function(x)  #a function that generates the estimates
> {
>    loglik<-function(alpha,theta1,theta2){
>        -(
> log(factorial(n)/(factorial(N1)*factorial(N2)))-r*lgamma(alpha)-alpha*N1*log(theta1)
> 
> -alpha*(r-N1)*log(theta2)+(alpha-1)*sum(log(cc[1:N1]))+(alpha-1)*sum(log(cc[w:r]-tau1+
> 
> (theta2/theta1)*tau1))-(1/theta1)*sum(cc[1:N1])-(1/theta2)*sum(cc[w:r]-tau1+(theta2/theta1)*tau1)+
> 
> (n-r)*log(1-pgamma((max(cc)-tau1+(theta2/theta1)*tau1)/theta2,alpha,1)))
>    }}
> 
> H<-hessian(m1, cc)


Is there a particular reason why you need the hessian?

PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, *reproducible* code.

Please make your function more readable.
We are happy to help, but that means you need to make your code readable 
and reproducible by giving some corresponding data.

Best wishes,
Uwe Ligges





> Thank you and take care.
> Laila
> 
> 	[[alternative HTML version deleted]]
> 
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.



More information about the R-help mailing list