scikit-learn · qinhanmin2014 · Oct 11, 2018 · Oct 1, 2018 · Oct 1, 2018 · Oct 1, 2018
diff --git a/doc/modules/classes.rst b/doc/modules/classes.rst
@@ -865,6 +865,7 @@ details.
    :template: function.rst
 
    metrics.explained_variance_score
+   metrics.max_error
    metrics.mean_absolute_error
    metrics.mean_squared_error
    metrics.mean_squared_log_error

diff --git a/doc/modules/model_evaluation.rst b/doc/modules/model_evaluation.rst
@@ -84,6 +84,7 @@ Scoring                           Function
 
 **Regression**
 'explained_variance'              :func:`metrics.explained_variance_score`
+'max_error'                       :func:`metrics.max_error`
 'neg_mean_absolute_error'         :func:`metrics.mean_absolute_error`
 'neg_mean_squared_error'          :func:`metrics.mean_squared_error`
 'neg_mean_squared_log_error'      :func:`metrics.mean_squared_log_error`
@@ -1530,6 +1531,38 @@ function::
     ... # doctest: +ELLIPSIS
     0.990...
 
+.. _max_error:
+
+Max error
+-------------------
+
+The :func:`max_error` function computes the maximum `residual error
+<https://en.wikipedia.org/wiki/Errors_and_residuals>`_ , a metric
+that captures the worst case error between the predicted value and
+the true value. In a perfectly fitted single output regression
+model, ``max_error`` would be ``0`` on the training set and though this
+would be highly unlikely in the real world, this metric shows the
+extent of error that the model had when it was fitted.
+
+
+If :math:`\hat{y}_i` is the predicted value of the :math:`i`-th sample,
+and :math:`y_i` is the corresponding true value, then the max error is
+defined as
+
+.. math::
+
+  \text{Max Error}(y, \hat{y}) = max(| y_i - \hat{y}_i |)
+
+Here is a small example of usage of the :func:`max_error` function::
+
+  >>> from sklearn.metrics import max_error
+  >>> y_true = [3, 2, 7, 1]
+  >>> y_pred = [9, 2, 7, 1]
+  >>> max_error(y_true, y_pred)
+  6
+
+The :func:`max_error` does not support multioutput.
+
 .. _mean_absolute_error:
 
 Mean absolute error

diff --git a/doc/whats_new/v0.21.rst b/doc/whats_new/v0.21.rst
@@ -52,6 +52,15 @@ Support for Python 3.4 and below has been officially dropped.
   graph, which would add explicitly zeros on the diagonal even when already
   present. :issue:`12105` by `Tom Dupre la Tour`_.
 
+
+:mod:`sklearn.metrics`
+......................
+
+- |Feature| Added the :func:`metrics.max_error` metric and a corresponding
+  ``'max_error'`` scorer for single output regression.
+  :issue:`12232` by :user:`Krishna Sangeeth <whiletruelearn>`.
+
+
 Multiple modules
 ................
 

diff --git a/sklearn/metrics/__init__.py b/sklearn/metrics/__init__.py
@@ -55,12 +55,14 @@
 from .pairwise import pairwise_distances_chunked
 
 from .regression import explained_variance_score
+from .regression import max_error
 from .regression import mean_absolute_error
 from .regression import mean_squared_error
 from .regression import mean_squared_log_error
 from .regression import median_absolute_error
 from .regression import r2_score
 
+
 from .scorer import check_scoring
 from .scorer import make_scorer
 from .scorer import SCORERS
@@ -99,6 +101,7 @@
     'log_loss',
     'make_scorer',
     'matthews_corrcoef',
+    'max_error',
     'mean_absolute_error',
     'mean_squared_error',
     'mean_squared_log_error',

diff --git a/sklearn/metrics/regression.py b/sklearn/metrics/regression.py
@@ -31,6 +31,7 @@
 
 
 __ALL__ = [
+    "max_error",
     "mean_absolute_error",
     "mean_squared_error",
     "mean_squared_log_error",
@@ -573,3 +574,37 @@ def r2_score(y_true, y_pred, sample_weight=None,
         avg_weights = multioutput
 
     return np.average(output_scores, weights=avg_weights)
+
+
+def max_error(y_true, y_pred):
+    """
+    max_error metric calculates the maximum residual error.
+
+    Read more in the :ref:`User Guide <max_error>`.
+
+    Parameters
+    ----------
+    y_true : array-like of shape = (n_samples)
+        Ground truth (correct) target values.
+
+    y_pred : array-like of shape = (n_samples)
+        Estimated target values.
+
+    Returns
+    -------
+    max_error : float
+        A positive floating point value (the best value is 0.0).
+
+    Examples
+    --------
+    >>> from sklearn.metrics import max_error
+    >>> y_true = [3, 2, 7, 1]
+    >>> y_pred = [4, 2, 7, 1]
+    >>> max_error(y_true, y_pred)
+    1
+    """
+    y_type, y_true, y_pred, _ = _check_reg_targets(y_true, y_pred, None)
+    check_consistent_length(y_true, y_pred)
+    if y_type == 'continuous-multioutput':
+        raise ValueError("Multioutput not supported in max_error")
+    return np.max(np.abs(y_true - y_pred))
diff --git a/sklearn/metrics/scorer.py b/sklearn/metrics/scorer.py
@@ -22,7 +22,7 @@
 
 import numpy as np
 
-from . import (r2_score, median_absolute_error, mean_absolute_error,
+from . import (r2_score, median_absolute_error, max_error, mean_absolute_error,
                mean_squared_error, mean_squared_log_error, accuracy_score,
                f1_score, roc_auc_score, average_precision_score,
                precision_score, recall_score, log_loss,
@@ -454,6 +454,8 @@ def make_scorer(score_func, greater_is_better=True, needs_proba=False,
 # Standard regression scores
 explained_variance_scorer = make_scorer(explained_variance_score)
 r2_scorer = make_scorer(r2_score)
+max_error_scorer = make_scorer(max_error,
+                               greater_is_better=False)
 neg_mean_squared_error_scorer = make_scorer(mean_squared_error,
                                             greater_is_better=False)
 neg_mean_squared_log_error_scorer = make_scorer(mean_squared_log_error,
@@ -498,6 +500,7 @@ def make_scorer(score_func, greater_is_better=True, needs_proba=False,
 
 SCORERS = dict(explained_variance=explained_variance_scorer,
                r2=r2_scorer,
+               max_error=max_error_scorer,
                neg_median_absolute_error=neg_median_absolute_error_scorer,
                neg_mean_absolute_error=neg_mean_absolute_error_scorer,
                neg_mean_squared_error=neg_mean_squared_error_scorer,

diff --git a/sklearn/metrics/tests/test_common.py b/sklearn/metrics/tests/test_common.py
@@ -40,6 +40,7 @@
 from sklearn.metrics import label_ranking_average_precision_score
 from sklearn.metrics import label_ranking_loss
 from sklearn.metrics import log_loss
+from sklearn.metrics import max_error
 from sklearn.metrics import matthews_corrcoef
 from sklearn.metrics import mean_absolute_error
 from sklearn.metrics import mean_squared_error
@@ -89,6 +90,7 @@
 #
 
 REGRESSION_METRICS = {
+    "max_error": max_error,
     "mean_absolute_error": mean_absolute_error,
     "mean_squared_error": mean_squared_error,
     "median_absolute_error": median_absolute_error,
@@ -399,7 +401,7 @@ def precision_recall_curve_padded_thresholds(*args, **kwargs):
     "micro_precision_score", "micro_recall_score",
 
     "matthews_corrcoef_score", "mean_absolute_error", "mean_squared_error",
-    "median_absolute_error",
+    "median_absolute_error", "max_error",
 
     "cohen_kappa_score",
 }
@@ -429,6 +431,7 @@ def precision_recall_curve_padded_thresholds(*args, **kwargs):
 # No Sample weight support
 METRICS_WITHOUT_SAMPLE_WEIGHT = {
     "median_absolute_error",
+    "max_error"
 }
 
 

diff --git a/sklearn/metrics/tests/test_regression.py b/sklearn/metrics/tests/test_regression.py
@@ -14,6 +14,7 @@
 from sklearn.metrics import mean_squared_error
 from sklearn.metrics import mean_squared_log_error
 from sklearn.metrics import median_absolute_error
+from sklearn.metrics import max_error
 from sklearn.metrics import r2_score
 
 from sklearn.metrics.regression import _check_reg_targets
@@ -29,6 +30,7 @@ def test_regression_metrics(n_samples=50):
                                            np.log(1 + y_pred)))
     assert_almost_equal(mean_absolute_error(y_true, y_pred), 1.)
     assert_almost_equal(median_absolute_error(y_true, y_pred), 1.)
+    assert_almost_equal(max_error(y_true, y_pred), 1.)
     assert_almost_equal(r2_score(y_true, y_pred),  0.995, 2)
     assert_almost_equal(explained_variance_score(y_true, y_pred), 1.)
 
@@ -59,6 +61,7 @@ def test_regression_metrics_at_limits():
     assert_almost_equal(mean_squared_log_error([0.], [0.]), 0.00, 2)
     assert_almost_equal(mean_absolute_error([0.], [0.]), 0.00, 2)
     assert_almost_equal(median_absolute_error([0.], [0.]), 0.00, 2)
+    assert_almost_equal(max_error([0.], [0.]), 0.00, 2)
     assert_almost_equal(explained_variance_score([0.], [0.]), 1.00, 2)
     assert_almost_equal(r2_score([0., 1], [0., 1]), 1.00, 2)
     assert_raises_regex(ValueError, "Mean Squared Logarithmic Error cannot be "

diff --git a/sklearn/metrics/tests/test_score_objects.py b/sklearn/metrics/tests/test_score_objects.py
@@ -47,7 +47,8 @@
                       'neg_mean_absolute_error', 'neg_mean_squared_error',
                       'neg_mean_squared_log_error',
                       'neg_median_absolute_error', 'mean_absolute_error',
-                      'mean_squared_error', 'median_absolute_error']
+                      'mean_squared_error', 'median_absolute_error',
+                      'max_error']
 
 CLF_SCORERS = ['accuracy', 'balanced_accuracy',
                'f1', 'f1_weighted', 'f1_macro', 'f1_micro',