@@ -345,8 +345,8 @@ def k_means(
345
345
centroid seeds. The final results will be the best output of
346
346
n_init consecutive runs in terms of inertia.
347
347
348
- When `n_init='auto'`, the number of runs will be 10 if using
349
- `init='random'`, and 1 if using `init='kmeans++'`.
348
+ When `n_init='auto'`, the number of runs depends on the value of init:
349
+ 10 if using `init='random'`, 1 if using `init='kmeans++'`.
350
350
351
351
.. versionadded:: 1.2
352
352
Added 'auto' option for `n_init`.
@@ -1208,8 +1208,8 @@ class KMeans(_BaseKMeans):
1208
1208
in terms of inertia. Several runs are recommended for sparse
1209
1209
high-dimensional problems (see :ref:`kmeans_sparse_high_dim`).
1210
1210
1211
- When `n_init='auto'`, the number of runs will be 10 if using
1212
- `init='random'`, and 1 if using `init='kmeans++'`.
1211
+ When `n_init='auto'`, the number of runs depends on the value of init:
1212
+ 10 if using `init='random'`, 1 if using `init='kmeans++'`.
1213
1213
1214
1214
.. versionadded:: 1.2
1215
1215
Added 'auto' option for `n_init`.
@@ -1736,8 +1736,8 @@ class MiniBatchKMeans(_BaseKMeans):
1736
1736
recommended for sparse high-dimensional problems (see
1737
1737
:ref:`kmeans_sparse_high_dim`).
1738
1738
1739
- When `n_init='auto'`, the number of runs will be 3 if using
1740
- `init='random'`, and 1 if using `init='kmeans++'`.
1739
+ When `n_init='auto'`, the number of runs depends on the value of init:
1740
+ 3 if using `init='random'`, 1 if using `init='kmeans++'`.
1741
1741
1742
1742
.. versionadded:: 1.2
1743
1743
Added 'auto' option for `n_init`.
0 commit comments