Should there be a validation set in the parameter optimization example?

Hi all,

In the parameter optimization loop examples (, the dataset is split into a training and test set. Shouldn’t there also be a validation set, since we are optimizing the parameters of the model?


Hi @Celestial,

you are right. The performance of the model should always be determined using a data set that was not used for training. So, you could simply do an additional partitioning of your data and then evaluate your model with the optimized parameters using the data set that was not used for the parameter optimization.

Let me know if you have further questions.



This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.