MNT Fix typos found by codespell (scikit-learn#31012)

Author: DimitriPapadopoulos

build_tools/codespell_ignore_words.txt

@@ -1,13 +1,16 @@
+achin
 aggresive
 aline
 ba
 basf
 boun
 bre
 cach
+chanel
 complies
 coo
 copys
+datas
 deine
 didi
 feld
@@ -17,11 +20,13 @@ fro
 fwe
 gool
 hart
+heping
 hist
 ines
 inout
 ist
 jaques
+lamas
 linke
 lod
 mape
@@ -31,6 +36,7 @@ nmae
 ocur
 pullrequest
 ro
+ser
 soler
 suh
 suprised
@@ -40,6 +46,8 @@ teh
 thi
 usal
 vie
+vor
 wan
+whis
 winn
 yau

doc/images/ml_map.svg

@@ -527,7 +527,7 @@ Some classification tasks can naturally exhibit rare classes: for instance,
 there could be orders of magnitude more negative observations than positive
 observations (e.g. medical screening, fraud detection, etc). As a result,
 cross-validation splitting can generate train or validation folds without any
-occurence of a particular class. This typically leads to undefined
+occurrence of a particular class. This typically leads to undefined
 classification metrics (e.g. ROC AUC), exceptions raised when attempting to
 call :term:`fit` or missing columns in the output of the `predict_proba` or
 `decision_function` methods of multiclass classifiers trained on different
@@ -903,7 +903,7 @@ is always used to train the model.
 
 This class can be used to cross-validate time series data samples
 that are observed at fixed time intervals. Indeed, the folds must
-represent the same duration, in order to have comparable metrics accross folds.
+represent the same duration, in order to have comparable metrics across folds.
 
 Example of 3-split time series cross-validation on a dataset with 6 samples::
 

doc/modules/cross_validation.rst

@@ -234,7 +234,7 @@ def plot(self, plot_method="contourf", ax=None, xlabel=None, ylabel=None, **kwar
                         cmap = "gist_rainbow"
 
                 # Special case for the tab10 and tab20 colormaps that encode a
-                # discret set of colors that are easily distinguishable
+                # discrete set of colors that are easily distinguishable
                 # contrary to other colormaps that are continuous.
                 if cmap == "tab10" and n_responses <= 10:
                     colors = plt.get_cmap("tab10", 10).colors[:n_responses]

sklearn/inspection/_plot/decision_boundary.py

@@ -1342,7 +1342,7 @@ def fake_score_func(y_true, y_pred):
     search_cv.set_params(scoring=fake_scorer)
     with pytest.warns(UserWarning, match="does not support sample_weight"):
         search_cv.fit(X, y, sample_weight=sw)
-    # multi-metric evalutation
+    # multi-metric evaluation
     search_cv.set_params(
         scoring=dict(fake=fake_scorer, accuracy="accuracy"), refit=False
     )

sklearn/model_selection/tests/test_search.py

@@ -318,7 +318,7 @@ def fit(self, X, y=None, sample_weight=None):
             )
 
         if self.strategy != "quantile" and sample_weight is not None:
-            # Preprare a mask to filter out zero-weight samples when extracting
+            # Prepare a mask to filter out zero-weight samples when extracting
             # the min and max values of each columns which are needed for the
             # "uniform" and "kmeans" strategies.
             nnz_weight_mask = sample_weight != 0