search

nipype / pydra-ml

Apache License 2.0
4 stars 10 forks source link

Change wording "split" to "bootstrap_samples" #17

Closed danielmlow closed 4 years ago

danielmlow commented 4 years ago

Just a thought:

Maybe people not used to using bootstrapping will think the pipeline is doing k-fold and will set n_splits to 5/7/10 which would lead to too few test sets for bootstrapping. "Split" sounds like k-fold non overlapping split, so I think the word "n_bootstrap_samples" will be more obvious.

If both k-fold and bootstrapping are available, then the word could be "n_samples"

satra commented 4 years ago

splits comes straight from the shufflesplit keywords. i would not change this as most people using scikit-learn understand what this is.