Difference between grid search node and parameter optimization loop

[06mickey]

Hello,
I want to know the difference between these two nodes : “Grid Search” &
“Parameter optimization loop” nodes . For what different works are they used
& what is the difference between them. Also, please share a video or workflow
in which cofiguration and use of Grid search node is explained.

[ScottF]

Hi @06mickey -

Are you referring to the GridSearch (3.6) node that’s part of the Weka extension? I’m not very familiar with it, but based on the node description it looks like it is automating some sort of cross validation in a regression context. We don’t have a video for this one since it’s based on Weka code.

The parameter optimization nodes are more generally applicable, and are what we recommend for sussing out the best combination of hyperparameters for a particular model and data set. We do have a video for these, as well as sample workflows on the Hub and even a component:

2 Examples for Parameter Optimization Loops – kathrin

Parameter Optimization – knime

In short, I wouldn’t bother with the Weka node - I’d focus on the others. Hope that helps.