Loop Cluster - exclude examples

aarapidi

Hi,
I have used Rapidminer to create many files with clusters. Therefore, my data file has many examples with a set of numeric attributes, a label, and a special attribute - cluster attribute. Next, I am trying to use (Loop Clusters) process to run classification models within validation process which is primarily working for the data files I have. However, some data files have clusters with one or 2 examples or observations; therefore an error is generated whenever the  loop reaches any of the clusters with small number of examples. My question is how can I exclude clusters with low number of examples in the Loop Clusters process.
Thanks

David_A

Hi,

you can extract the size of your cluster subset with the Extract Macro operator and the option "number of examples", next you use a branch operator, set as condition that the number of examples has to be greater two, and perform your model building in the Then-branch. The Else-branch can return an empty example set or some dummy data, whatever fits best.

Best,
David

aarapidi

Thank you!