"[SOLVED] k-means parameters"

MarcosRL

Hello friends of the community a question.
I am studying the theory of the algorithm "k-means" to understand its inner workings and I have some concerns with some parameters that appear in rapid miner.
The parameter "determine good start values" what applies mathematical rule?
The parameter "Local use randon seed" what is the role?
The parameter "Max optimization steps" what applies mathematical rule?
regards

MarcosRL

any ideas ?
Regards

Unknown

Hello

"determine good start values" - if you hover your mouse pointer over the parameter you will get some additional text.
"Local use randon seed" - this sets the random seed to a specific value so that all random numbers are generated in a reproducible way. The k-means algorithm chooses the initial cluster prototypes at random.
"Max optimization steps" - this dictates the maximum number of steps the k-means algorithm should take before stopping as it tries to find the best cluster prototypes within a run (which is the max runs parameter)

regards

Andrew

MarcosRL

thanks