Model Simulator functionality

[ak288]

Hi everyone,

I’m new to KNIME and worked with Rapidminer before.

In Rapidminer is used to use the operator/node “Model Simulator”. It gives you the opportunity to change the input data of a model and shows as an output predictions, confidences, and explanations for those inputs. The whole description of the Rapidminer operator is described on the Rapidminer website Rapidminer Model Simulator

My question now is, is there a similar node in KNIME availiable or does anyone have a sugesstion how to implement something like this within KNIME ?

Thanks in advance!

[ScottF]

Hi there -

Sounds like you might be interested in some of our components for model intepretability. This recent blog post touches on several of these components and provides example workflows that shows how they work in context:

Understand Your ML Model with Global Feature Importance

If this isn’t quite what you’re looking for, reply here and we can dig a bit deeper.

[ak288]

Thanks Scott for your reply. The link you posted is really interesting and helps me a lot.

But I am looking for a quite different function.
The function I am looking for allows you to enter the desired output and gives you then back the needed input data to achieve this output.

Is there a opportunity in Knime to implement such a simulation?

[Daniel_Weikert]

That sounds like Solver or Goal Seek in Excel. I have not seen this so far but if someone wants to enlighten me.