Matlab Bootstrap not complete resampling

Question

I would like to perform bootstrap in Matlab. I have 100 original data points and I would like each iteration of bootstrap to choose only 57 points with replacement randomly. How do I accomplish it?

I cannot seem to find this functionality in Matlab function bootstrp.

Regards,

Answer 1

To choose n points out of a vector randomly with replacement: use randi to generate the (possibly repeated) indices:

vector = (1:100).^2; %// example data
n = 57;
ind = randi(numel(vector),1,n); %// n random integers between 1 and numel(vector)
sample = vector(ind);

To do it directly with bootstrp: let fun denote the function you would pass to bootstrp. You just need to pick the first 57 values of each 100-value sample:

vector = (1:100).^2; %// example data
n = 57;
nboot = 10;
bootstrp(nboot, @(x) fun(x(1:57)), vector)