I don't know how to set forty centers as in discribed as below, so i nee d help with it please!

 1 partitions 20000 input points into 40 clusters. The dimensions of the data being tested are 24, 32, and 40, respectively. In Experiment 1, input points were generated as follows: *for each of the mentioned dimension, set the forty centers as {(⎯1, –1, …, –1), (–0.95, –0.95, …, –0.95), (–0.9, ⎯0.9, …, –0.9), (–0.85, –0.85, …, –0.85), (0.9, 0.9, …, 0.9), (0.95, 0.95, …, 0.95)}.* Then, use Gaussian distribution with mean μ = 0 and standard deviation σ = 0.05 to generate 500 points around each center. The 40 initial centers to run Stage I of the Kmeans are chosen randomly from data points.

Answer 1

[Walter Roberson]

Supposing you have a num_samples by num_measurements array of coordinates, then

   trial_centers = repmat((-1:.05:0.95).', 1, num_measurements);

You can now proceed to generate 500 points around each center. That will get you 500 * 40 = 20,000 multi-dimensional points.

Now out of those 20,000 multi-dimensional points, select 40 of them at random and use those 40 as the initial centers for kmeans.

[Fereshteh....]

Thank you very much for your answer, actually I think I should put the mean of each Gaussian function on each center to make 500 points around them, I want to use obj = gmdistribution(mu,sigma,p); you think it is correct? I have never worked with matlab for this kind of problems so I am clueless, sorry!

[Walter Roberson]

That sounds plausible.