Scoring different neural networks in a loop

KNIME Analytics Platform
1

Dear all,

I would like to test optimal settings for a neural network (max nr of iterations, nr of hidden layers, nr of hidden neurons) for my dataset. For that purpose, I have made a table of these settings and used the 'table row to variable loop start' to train several neural networks. Now, I would like to apply the 'numeric scorer' to each of the iteration (12 in total). How should I do that?

Best regards,

Frits

2

Put the scorer before the apropriate Loop End node, that should do what you want.

3

Thanks for your answer. If I put the scorer before the Loop End node, I only get the  scoring results and not the table with predicted results. I have solved it by making a separate loop after the training loop to do the scoring.

4

You could use a loop end 2 ports to capture both the scoring table and the predictions?