I understand that e-means does the same thing as k-means but faster, that's great, congratulations on that. I'm just saying that the example you've chosen to showcase this result is terrible. The goal of a clustering algorithm is to discover the structure of the data and in your example the structure is simply not there (to be precise, there is only a bit of structure but you're trying to find a lot more of it than there is). I think it would be much more convincing if you showed an example where the number of clusters you're looking for actually matches the number of clusters in the data.