How to extract predicted values from regression and XGBoost models

[Subha_D]

Hello there,

I have built a couple of models to predict IC50 (drug response) values and I would like to extract the predicted value for each type of cell line and drug in my dataset. How would I go about doing that?

My dataset looks like this: there are IC50 values for a certain drug and it’s response on a certain cell line, each unique combination make up the rows. And I would like to know the predicted value for each of these combinations.
How would I extract in a table the unique row ID and associated prediction?

Any help would be appreciated.

Thanks,
Subha

[aworker]

Hi @Subha_D & welcome back to the KNIME forum,

Maybe the following workflow by @deepsika might answer your question to start with:

Bioactivity Prediction using ML – deepsika

The following version is updated and metanodes have been converted to components:

20230621 Bioactivity prediction 18081.knwf (2.0 MB)

This workflow is adressing the problem of biological activity so I guess the closest to your IC50 prediction problem.

Hope it helps.

Best

Ael

[Subha_D]

Hi @aworker !

Thanks you for your reply!

[aworker]

Hi @Subha_D

My pleasure !

Did the workflow in the reply solved your question ? If so, could you please check it as the solution so that other people can find it easily ? Thank you.

All the best,
Ael