Is there any way to reapply a preprocessment done in a training dataset to a new dataset of experimental data for submission of the transformed data to the already trained classifier?
The preprocessor modifies the domain
on the training data set. If you want to apply the same transformations on the testing (experimental) data, you apparently have to cast it into the same domain, as the Orange's built-in predictors seem to do:
train = preprocess(train)
test = Table(train.domain, test)