Hallo,
I'm working with the mle function and I would like to ask you a couple
of questions.
My goal is to construct the historical value of v1(t), v2(t) and
v3(t) using the maximum likelihood estimation.
So, I need to optimize the following log-likelihood:
sum(E1_f[t,]*(v1*teta1[] + v2*teta2[] + v3*teta3[]) - E_f[t,]*log(1 +
exp(v1*teta1[] + v2*teta2[] + v3*teta3[])))
(E_f and E1_f are 136x111 matrices and teta1,teta2 and teta3 are
111x1 vectors).
By writing the code below, I just obtain the result for t=1.
################
library(stats4)
likelihood <- function(v1,v2,v3){
for (t in 1:136){
return(-(sum(E1_f[t,]*(v1*teta1[] + v2*teta2[] + v3*teta3[]) -
E_f[t,]*log(1 + exp(v1*teta1[] + v2*teta2[] + v3*teta3[])))))
}
}
L_f <- mle(minuslog=likelihood, start=list(v1=1, v2=2, v3=3))
x <- summary(L_f)
################
What should I change in the code?
And how can I store the values of v1(t), v2(t) and v3(t) in 3 vectors,
in order to use them after?
Thank you very much.
Roger
