Displaying 1 result from an estimated 1 matches for "newdis1".
Did you mean:
newdisk
2012 Jul 11
0
declaring negative log likelihood of a distribution
...frequency counts of the above
> binomial values, totally I have 399 samples
*
Now to estimate the parameters, I first declared the negative log
likelihood function of this new distribution as below(I used two methods,
one with a loop and one without the loop)
#without any loops
*> Loglik.newdis1<- function(x,a,b,fre,n){
KBLL1<-sum(fre*(log(a*b)+lchoose(n,x)+log(sum((-1)**(0:1000) *
choose(b-1, 0:1000) * beta(x+a+a*0:1000, n-x+1)))))
return(-KBLL1)
}*
#here a and b are the parameters to be estimated and x=binomial
values, fre=frequencies and n=binomial trials...