Could anyone help me how to find the seed of the centroid in a random manner.

The following code executes and gives the result.

code:

clear all

clc

centroids=2

dimensions = 2;

iterations = 10;

dataset_subset = 2;

load test1.mat

meas = meas(:,1+dataset_subset:dimensions+dataset_subset);

dataset_size = size (meas);

[idx,KMEANS_CENTROIDS] = kmeans(meas,centroids);

With respect to the above code the centroids is fixed to 2.

Could anyone help me to modify the code in such a way that the number of centroids needs to be selected in a random manner.