scikit-learn-mooc
scikit-learn-mooc copied to clipboard
Dataset transition between adult_census (Predictive Modeling Pipeline) and boston_housing (Select the Best Model) and then back to adult_census (HyperParameter Tuning)
Now that the hyper-parameter tuning has been moved out of the modelling predictive pipeline part, we do:
- adult_census for Modeling Predictive Pipeline
- boston_housing for Selecting the Best Model
- back to adult_census for Hyperparameter tuning
do we care?
Can we use adult_census for Selecting the Best Model?
Can we use adult_census for Selecting the Best Model?
We should try, it would probably help streamlining the teaching.
I think that this is actually useful to have a regression dataset for the module on the evaluation. It allows putting in perspective the variation of the target in a natural unit. It might not be so easy to have similar intuitions with classification.
Let's close this as not important enough.