evaluattion

nguyenxuanhau

I run  EMClustering (of Rapid) and W_EM algorithms (of Weka which intergrated into Rapid) with the same value parameters on the same data but the result of two algorithms is different. Why is it?
Help me
Thanks

land

Hi,
because it are two different implementations of an idea of an algorithm.

Don't want to good in more detail, but probably the (random) initialization is performed differently.

Greetings,
Sebastian