RandomForestRegressor doesn't accept max_samples=1.0

Describe the bug

This example from the doc works:

from sklearn.ensemble import RandomForestRegressor
from sklearn.datasets import make_regression
X, y = make_regression(n_features=4, n_informative=2,
                       random_state=0, shuffle=False)
regr = RandomForestRegressor(max_depth=2, random_state=0)
regr.fit(X, y)
print(regr.predict([[0, 0, 0, 0]]))

Just changing one line to this:

regr = RandomForestRegressor(max_depth=2, random_state=0, max_samples=1.0)

doesn’t work anymore:

ValueError: `max_samples` must be in range (0, 1) but got value 1.0

I believe max_samples=None (the default) and max_samples=1.0 should behave the same.

Steps/Code to Reproduce

see above

Versions

System:
    python: 3.8.5 (default, Jan 27 2021, 15:41:15)  [GCC 9.3.0]
executable: /usr/bin/python3
   machine: Linux-5.8.0-53-generic-x86_64-with-glibc2.29

Python dependencies:
          pip: 20.0.2
   setuptools: 45.2.0
      sklearn: 0.24.2
        numpy: 1.17.4
        scipy: 1.6.3
       Cython: None
       pandas: None
   matplotlib: 3.4.2
       joblib: 1.0.1
threadpoolctl: 2.1.0
Built with OpenMP: True

[UnixJunkie]

max_samples=0.0 should crash (you cannot train a model without training samples), max_samples=1.0 should work just fine (you can train a model using all the available training samples)

On Fri, May 28, 2021 at 3:02 PM murata-yu @.***> wrote:

Yes, max_samples=1.0 and max_samples=None should behave the same . But, max_samples is not expected to be greater equal than 1.0 . So, I think it is correct tha an error occured at max_samples=1.0 .

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/scikit-learn/scikit-learn/issues/20156#issuecomment-850159999, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAFUFABWJVGAEGHFE7TXDNTTP4WWRANCNFSM45VVMK4Q .

[UnixJunkie]

Well spotted, please send a PR.