machine learning - How to apply InformationGain in rapidminer with seperate test set ? -

I am working with text classification in the tower rapidly. I have separate tests and training divisions. I am using n-fold cross validation to get information in a dataset, but how am I confused to implement it on a separate test set? The image below

I added the word list output to the shape First is the use of "Process Documents from Files" which is used for training for "processed documents from files" which is used for testing but I want to apply the lesser feature to "Process Documents from Files" Which should probably come back. "Select from the weight" (lower dimension) from the operator, but it returns the weight, which I can not provide second to "process documents from files". I searched a lot but managed to find anything that could satisfy my needs?

Is the Rapidiner split different tests / trains and is it really possible to implement a feature selection?

Is there any way to convert these weight to the word list? Please do not say that I will write in the repository (I can not do this)?

In such a scenario, when I have a separate examination / train, and I need to apply the feature selection, how will I ensure that the exam /

model immediately after the lower process document operator Before Apply , enter a new select from Weight operator. Use the Multiplication operator to copy the weight from the operator and select it from the new Weight operator. .