Hello,
i’ve found that after performing an optimization loop the best parameters selected are not passed back to the learner and then eventually it’s necessary to set them manually after the optimization loop is concluded.
Is this right? If so is there an efficient way to pass the parameter without “copy and paste” them?
Hi @wincenzo and welcome the KNIME Forum
After the loops are finished, the best parameter are available in the Parameter Optimization Loop End (upper port). With a Table Row to Variable node you can create flow variables for your best parameters. These flow vars are input for your Learner Node. Is this where you are looking for?
gr. Hans
hi @HansS thank for the reply, yes this is what i’m looking for, and if i got it right it seems that i have to create another twin flow “learner-predictor” to pass the optimal parameter, correct?
Yep, with a data partition that is not used to optimize your parameters.
with a data partition that is not used to optimize your parameters.
What do you mean? Sorry if i insist.
yes sure i know about train/test did you just mean that i have to perform prediction on test set?
To my opinion, build your model with a dataset in 3 partitions
train: optimize your parameters
validate: build your model on, with the optimized parameters
test: final model-performing statistics
yes you are right, thanks for your help.
This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.