bootstrap confidence interval for random samples

Question

I have a list of 500 data (y=500)
I am using bootstrap method in matlab in order to calculate confidence interval.
I am using bootci function,
bootci(1000,@mean,randsample(y, 50, true))
Normally: Here the 50 random data is re-sampled(with replacement) 1000 times from the same 50 data.
I want: Can i do something so the 50 data will be re-sampled 1000 times from not fixed 50 random data but from 500(entire) data?

Is there any other function helps to do it? any solution please?

Answer 1

Is this what you are looking for?:

bootci(1000, @(x) mean(randsample(x, 50, true)), y)

By moving the resampling inside the bootfun you will get a (potentially) new 50-element sample during each bootsrap sampling (1000 times). Furthermore, by specifying y as the data argument for bootci you achieve that the 500-element (full) data is sampled.