Hell,
I am using Maxboot to compute the bootstrapped estimates of parameters of a maximum likelihood function, but I want to retrieve all the random samples that were used to compute these estimates.
Any help would be appreciated.
Thanks.
Sincerely,
Sajjadur.
1 Answer
0
With some programming it would only be possible to retrieve the number of times each observation occurred in the re-sampling.
If it is important to look at each sample in the re-sampling, put the call to Maxlik in a loop and generate the bootstrap samples yourself using the GAUSS function exctsmpl():
x0 = {};
for I(1,100,1);
subsample = exctsmpl("originalData","newData",100);
{ x,f,g,cov,ret } = maxlik("newData",0,&lpr,x0);
if ret == 0;
x0 = x0 | x';
endif;
endfor;
print "bootstrap estimate";
print meanc(x0);
print "standard errors";
print sqrt(diag(varcovx(x0)));
Your Answer
1 Answer
With some programming it would only be possible to retrieve the number of times each observation occurred in the re-sampling.
If it is important to look at each sample in the re-sampling, put the call to Maxlik in a loop and generate the bootstrap samples yourself using the GAUSS function exctsmpl():
x0 = {};
for I(1,100,1);
subsample = exctsmpl("originalData","newData",100);
{ x,f,g,cov,ret } = maxlik("newData",0,&lpr,x0);
if ret == 0;
x0 = x0 | x';
endif;
endfor;
print "bootstrap estimate";
print meanc(x0);
print "standard errors";
print sqrt(diag(varcovx(x0)));