Default `random_state` for `KMeans` and `FuzzyCMeans` should be `None`
vnmabus opened this issue · comments
Bug description summary
The default behavior of KMeans
and FuzzyCMeans
when no random_state
is passed should be to not use a fixed seed, as it is done in scikit-learn. Somehow, the current behavior is having a fixed seed (0).
Code to reproduce the bug
from skfda.ml.clustering import KMeans
clusterer = KMeans()
print(clusterer.random_state)
Expected result
None
should be printed.
Actual result
A fixed seed of 0
is printed.
Traceback (if an exception is raised)
No response
Software versions
scikit-fda version: 0.9.2.dev0
Additional context
No response