Logistic Regression - Error Execute failed: The logisitic regression model cannot be computed.

KNIME Analytics Platform
1

Hi I'm trying to produce a simple logistic regression but it keeps giving an error that I can't figure how to get rid off.

If I try with the same data in R it works just fine!

I suspect this is related to a inherent problem with the Knime operator but would appreciate if you can validate that.

I provide attached the workflow, the data and the log.

Thanks a lot for all your help!

 

Kind regards,

Joel

2

Hi raluar,

After some fiddling, I found that the "reference category" setting is the culprit for the error in your case. If you set it to 1 or -1 instead of NA, the model runs. You probably want to set it to -1 to have your resulting model coefficients with regards to 1.

This thread may also prove useful: https://tech.knime.org/forum/knime-general/logistic-regression

Note: this applies to the case when the domain is defined (as is in this example). If this is not the case: " The reference category is empty if the domain of the target column is not available. In this case the node determines the domain values right before computing the logistic regression model and chooses the last domain value as the targets reference category. "

3

Hello Kagaherk, thanks a lot. I overlooked that option.

 

Kind regards, Joel