fix compatibility with older scikit-learn

Author: jameslamb

dask_ml/metrics/regression.py

@@ -81,14 +81,50 @@ def mean_absolute_error(
     return result
 
 
-@derived_from(sklearn.metrics)
 def mean_absolute_percentage_error(
     y_true: ArrayLike,
     y_pred: ArrayLike,
     sample_weight: Optional[ArrayLike] = None,
     multioutput: Optional[str] = "uniform_average",
     compute: bool = True,
 ) -> ArrayLike:
+    """Mean absolute percentage error regression loss.
+
+    Note here that we do not represent the output as a percentage in range
+    [0, 100]. Instead, we represent it in range [0, 1/eps]. Read more in
+    https://scikit-learn.org/stable/modules/model_evaluation.html#mean-absolute-percentage-error
+
+    Parameters
+    ----------
+    y_true : array-like of shape (n_samples,) or (n_samples, n_outputs)
+        Ground truth (correct) target values.
+    y_pred : array-like of shape (n_samples,) or (n_samples, n_outputs)
+        Estimated target values.
+    sample_weight : array-like of shape (n_samples,), default=None
+        Sample weights.
+    multioutput : {'raw_values', 'uniform_average'} or array-like
+        Defines aggregating of multiple output values.
+        Array-like value defines weights used to average errors.
+        If input is list then the shape must be (n_outputs,).
+        'raw_values' :
+            Returns a full set of errors in case of multioutput input.
+        'uniform_average' :
+            Errors of all outputs are averaged with uniform weight.
+    compute : bool
+        Whether to compute this result (default ``True``)
+
+    Returns
+    -------
+    loss : float or array-like of floats in the range [0, 1/eps]
+        If multioutput is 'raw_values', then mean absolute percentage error
+        is returned for each output separately.
+        If multioutput is 'uniform_average' or ``None``, then the
+        equally-weighted average of all output errors is returned.
+        MAPE output is non-negative floating point. The best value is 0.0.
+        But note the fact that bad predictions can lead to arbitarily large
+        MAPE values, especially if some y_true values are very close to zero.
+        Note that we return a large value instead of `inf` when y_true is zero.
+    """
     _check_sample_weight(sample_weight)
     epsilon = np.finfo(np.float64).eps
     mape = abs(y_pred - y_true) / da.maximum(y_true, epsilon)

docs/source/modules/api.rst

@@ -245,6 +245,7 @@ Regression Metrics
    :toctree: generated/
 
    metrics.mean_absolute_error
+   metrics.mean_absolute_percentage_error
    metrics.mean_squared_error
    metrics.mean_squared_log_error
    metrics.r2_score

tests/metrics/test_regression.py

@@ -5,22 +5,28 @@
 import sklearn.metrics
 
 import dask_ml.metrics
+from dask_ml._compat import SK_024
+
+_METRICS_TO_TEST = [
+    "mean_squared_error",
+    "mean_absolute_error",
+    "r2_score",
+]
+
+# mean_absolute_percentage_error() was added in scikit-learn 0.24.0
+if SK_024:
+    _METRICS_TO_TEST.append("mean_absolute_percentage_error")
 
 
 @pytest.fixture(
-    params=[
-        "mean_squared_error",
-        "mean_absolute_error",
-        "mean_absolute_percentage_error",
-        "r2_score",
-    ]
+    params=_METRICS_TO_TEST
 )
 def metric_pairs(request):
     """Pairs of (dask-ml, sklearn) regression metrics.
 
     * mean_squared_error
     * mean_absolute_error
-    * mean_absolute_percentage_error
+    * mean_absolute_percentage_error (if scikit-learn >= 0.24.0)
     * r2_score
     """
     return (