[SOLVED] How to remove features that appear rarely in training data?

Flash

Hello,

I am using RapidMiner to classify some documents. Currently I want to apply a very simple feature selection technique, but I cannot find a method to do it.  ???

I just want to remove all attributes that appear less than four times in the trainingdata. Does anyone know which operator I will need?

Thanks for the help.  8)

Flash

I found a solution that can be used with text processing tool kit. The operators which create the word vector are able to "prune". If the value there is set to absolute something similar to the thing I asked for can be done.