Problem using Parameter Optimization loop and JPMML Regressor Predictor

[glimachave]

Hello. I’m trying to run an optimization using a model that I’ve created. The idea is to create a table with a single row containing the parameters that I want to optimize, and these parameters in this column will be controlled by the optimization.

The problem is when I try to use the variables from the optimization to set the table variables in Table Creator, I get the error

So apparently, I don’t know why but the numbers that are created by the optimization are being read as strings, in spite of having set the table to double.

Here is the workflow:

[nemad]

Hello @glimachave,

you might want to use the Variable to Table Row node instead of the Table Creator here.

Cheers,

Adrian

[glimachave]

Hi nemad,

I’ve tried to do so, but the error still persists.

[nemad]

What types do the columns output by the Variable to Table Row have?

[glimachave]

Int and Double. I’ve solved the problem using the Regression Predictor node, as the model is a polynomial regression. But it still doesn’t work if I intend to use the JPMML Regression Predictor.

[nemad]

Hi,

I tried to reproduce the problem but without success.
What kind of model are you using? Was it build with KNIME Analytics Platform?

Cheers,

Adrian