Uncertainty in Random Forest/SVM Models

Dear Knimers,

I am working on a binary classification random forest model to identify active and inactive ligands. I am wondering if it is possible to calculate uncertainty in Knime using random forest or SVM models?

Thanks

Hey @naraj,

have a look at the Entropy Uncertainty Scorer. In case you are using the uncertainty score for an active learning approach, you can also have a look at this simple workflow using and SVM model: Active Learning with Basic SVM Model

I hope that helps. Let me know if you have more questions,
Janina

2 Likes

Hi Janina,

Thanks for your reply.
I am wondering if I can use normalised class distributions obtained from SVM model as an input for Entropy Uncertainty Scorer node?

Best wishes,
Nasir

Hi @naraj,

yes, you can simply use the normalised class distributions from the SVM predictor node as an input.

Kind regards,
Janina

2 Likes

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.