TimeSeriesSplit and required minimum of n_splits #11893
Description
Description
TimeSeriesSplit.split() throws error for n_splits=1 even though in documentation it requires at least 1. Not sure if documentation needs to be updated or bug in TimeSeriesSplit?
Steps/Code to Reproduce
Example:
timeseries= np.arange(0,10)
tscv = TimeSeriesSplit(n_splits=1)
for train, test in tscv.split(timeseries):
print(train, test)
Expected Results
No error is thrown. In the documentation it requires n_splits at least 1 (and not at least 2).
Actual Results
------------------------------------------------------------------------
ValueError Traceback (most recent call last)
<ipython-input-180-b869ddb34868> in <module>()
1 time = np.arange(0,10)
----> 2 tscv = TimeSeriesSplit(n_splits=1)
3 for train, test in tscv.split(ts):
4 print(train, test)
~/.conda/envs/egym/lib/python3.6/site-packages/sklearn/model_selection/_split.py in __init__(self, n_splits, max_train_size)
725 super(TimeSeriesSplit, self).__init__(n_splits,
726 shuffle=False,
--> 727 random_state=None)
728 self.max_train_size = max_train_size
729
~/.conda/envs/egym/lib/python3.6/site-packages/sklearn/model_selection/_split.py in __init__(self, n_splits, shuffle, random_state)
282 "k-fold cross-validation requires at least one"
283 " train/test split by setting n_splits=2 or more,"
--> 284 " got n_splits={0}.".format(n_splits))
285
286 if not isinstance(shuffle, bool):
ValueError: k-fold cross-validation requires at least one train/test split by setting n_splits=2 or more, got n_splits=1.
Versions
Linux-3.10.0-693.2.2.el7.x86_64-x86_64-with-centos-7.4.1708-Core
Python 3.6.5 | packaged by conda-forge | (default, Apr 6 2018, 13:39:56)
[GCC 4.8.2 20140120 (Red Hat 4.8.2-15)]
NumPy 1.14.3
SciPy 1.1.0
Scikit-Learn 0.19.2