scikit-learn · Bhavya1705 · Aug 27, 2021 · Aug 27, 2021 · Aug 27, 2021 · Aug 27, 2021
diff --git a/sklearn/metrics/__init__.py b/sklearn/metrics/__init__.py
@@ -35,6 +35,10 @@
 from ._classification import zero_one_loss
 from ._classification import brier_score_loss
 from ._classification import multilabel_confusion_matrix
+from ._classification import sensitivity
+from ._classification import specificity
+from ._classification import miss_rate
+from ._classification import fallout_rate
 
 from . import cluster
 from .cluster import adjusted_mutual_info_score
@@ -116,6 +120,7 @@
     "euclidean_distances",
     "explained_variance_score",
     "f1_score",
+    "fallout_rate",
     "fbeta_score",
     "fowlkes_mallows_score",
     "get_scorer",
@@ -140,6 +145,7 @@
     "mean_tweedie_deviance",
     "median_absolute_error",
     "mean_absolute_percentage_error",
+    "miss_rate",
     "multilabel_confusion_matrix",
     "mutual_info_score",
     "ndcg_score",
@@ -167,6 +173,8 @@
     "SCORERS",
     "silhouette_samples",
     "silhouette_score",
+    "sensitivity",
+    "specificity",
     "top_k_accuracy_score",
     "v_measure_score",
     "zero_one_loss",

diff --git a/sklearn/metrics/_classification.py b/sklearn/metrics/_classification.py
@@ -20,6 +20,8 @@
 #          Bernardo Stein <bernardovstein@gmail.com>
 #          Shangwu Yao <shangwuyao@gmail.com>
 #          Michal Karbownik <michakarbownik@gmail.com>
+#          Bhavya Bhardwaj <bhavya1705@yahoo.com>
+# 		   Rohit Mittal <rohitm132@gmail.com>
 # License: BSD 3 clause
 
 
@@ -2649,3 +2651,207 @@ def brier_score_loss(y_true, y_prob, *, sample_weight=None, pos_label=None):
             raise
     y_true = np.array(y_true == pos_label, int)
     return np.average((y_true - y_prob) ** 2, weights=sample_weight)
+
+
+def fallout_rate(y_true, y_pred):
+    """Compute miss rate of a classification.
+
+    By definition the miss rate of a classification is defined as ``fp / (fp + tn)``,
+    where ``tn`` is the number of true negatives and ``fp`` the number of false
+    positives.
+
+    Parameters
+    ----------
+    y_true : array-like of shape (n_samples,)
+        Ground truth (correct) target values.
+
+    y_pred : array-like of shape (n_samples,)
+        Estimated targets as returned by a classifier.
+
+    Returns
+    -------
+    miss rate: The true positive rate (float)
+
+    References
+    ----------
+    .. [1] https://www.split.io/glossary/false-positive-rate/
+
+    Examples
+    --------
+    >>> from sklearn.metrics import sensitivity
+    >>> y_true = [0, 1, 0, 1]
+    >>> y_pred = [1, 1, 1, 0]
+    >>> fallout_rate(y_true, y_pred)
+    1.0
+    """
+    average = (None,)
+    warn_for = "Fall out"
+    zero_division = "warn"
+    y_true = column_or_1d(y_true)
+    y_pred = column_or_1d(y_pred)
+    assert_all_finite(y_true)
+    assert_all_finite(y_pred)
+    tn, fp, fn, tp = confusion_matrix(y_true, y_pred).ravel()
+    fall_out = _prf_divide(
+        np.array([fp]),
+        np.array([fp + tn]),
+        "Fall Out",
+        "predicted",
+        average,
+        warn_for,
+        zero_division,
+    )
+    return fall_out[0]
+
+
+def miss_rate(y_true, y_pred):
+    """Compute miss rate of a classification.
+
+    By definition the miss rate of a classification is defined as ``fn / (tp + fn)``,
+    where ``tp`` is the number of true positives and ``fn`` the number of
+    false negatives.
+
+    Parameters
+    ----------
+    y_true : array-like of shape (n_samples,)
+        Ground truth (correct) target values.
+
+    y_pred : array-like of shape (n_samples,)
+        Estimated targets as returned by a classifier.
+
+    Returns
+    -------
+    miss rate: The true positive rate (float)
+
+    References
+    ----------
+    .. [1] https://www.split.io/glossary/false-positive-rate/
+
+    Examples
+    --------
+    >>> from sklearn.metrics import sensitivity
+    >>> y_true = [0, 1, 0, 1]
+    >>> y_pred = [1, 1, 1, 0]
+    >>> miss_rate(y_true, y_pred)
+    0.5
+    """
+    average = (None,)
+    warn_for = "Miss rate"
+    zero_division = "warn"
+    y_true = column_or_1d(y_true)
+    y_pred = column_or_1d(y_pred)
+    assert_all_finite(y_true)
+    assert_all_finite(y_pred)
+    tn, fp, fn, tp = confusion_matrix(y_true, y_pred).ravel()
+    miss_rate = _prf_divide(
+        np.array([fn]),
+        np.array([fn + tp]),
+        "Miss Rate",
+        "predicted",
+        average,
+        warn_for,
+        zero_division,
+    )
+    return miss_rate[0]
+
+
+def specificity(y_true, y_pred):
+    """Compute specificity of a classification.
+
+    By definition the specificity of a classification is defined as ``tn / (tn + fp)``,
+    where ``tn`` is the number of true negatives and ``fp`` the number of false
+    positives.
+
+    Parameters
+        ----------
+    y_true : array-like of shape (n_samples,)
+        Ground truth (correct) target values.
+
+    y_pred : array-like of shape (n_samples,)
+        Estimated targets as returned by a classifier.
+
+    Returns
+    -------
+    specificity : The true positive rate (float)
+
+    References
+    ----------
+    .. [1] https://www.split.io/glossary/false-positive-rate/
+
+    Examples
+    --------
+    >>> from sklearn.metrics import sensitivity
+    >>> y_true = [0, 1, 0, 1]
+    >>> y_pred = [1, 1, 1, 0]
+    >>> specificity(y_true, y_pred)
+    0.0
+    """
+    average = (None,)
+    warn_for = "specificity"
+    zero_division = "warn"
+    y_true = column_or_1d(y_true)
+    y_pred = column_or_1d(y_pred)
+    assert_all_finite(y_true)
+    assert_all_finite(y_pred)
+    tn, fp, fn, tp = confusion_matrix(y_true, y_pred).ravel()
+    specificity = _prf_divide(
+        np.array([tn]),
+        np.array([tn + fp]),
+        "Specificity",
+        "predicted",
+        average,
+        warn_for,
+        zero_division,
+    )
+    return specificity[0]
+
+
+def sensitivity(y_true, y_pred):
+    """Compute sensitivity of a classification.
+
+    By definition the sensitivity of a classification is defined as ``tp / (tp + fn)``,
+    where ``tp`` is the number of true positives and ``fn`` the number of false
+    negatives.
+
+    Parameters
+    ----------
+    y_true : array-like of shape (n_samples,)
+        Ground truth (correct) target values.
+
+    y_pred : array-like of shape (n_samples,)
+        Estimated targets as returned by a classifier.
+
+    Returns
+    -------
+    sensitivity : The true positive rate (float)
+
+    References
+    ----------
+    .. [1] https://www.split.io/glossary/false-positive-rate/
+
+    Examples
+    --------
+    >>> from sklearn.metrics import sensitivity
+    >>> y_true = [0, 1, 0, 1]
+    >>> y_pred = [1, 1, 1, 0]
+    >>> sensitivity(y_true, y_pred)
+    0.5
+    """
+    average = (None,)
+    warn_for = "senstivity"
+    zero_division = "warn"
+    y_true = column_or_1d(y_true)
+    y_pred = column_or_1d(y_pred)
+    assert_all_finite(y_true)
+    assert_all_finite(y_pred)
+    tn, fp, fn, tp = confusion_matrix(y_true, y_pred).ravel()
+    sensitivity = _prf_divide(
+        np.array([tp]),
+        np.array([tp + fn]),
+        "Specificity",
+        "predicted",
+        average,
+        warn_for,
+        zero_division,
+    )
+    return sensitivity[0]
diff --git a/sklearn/metrics/tests/test_classification.py b/sklearn/metrics/tests/test_classification.py
@@ -42,6 +42,10 @@
 from sklearn.metrics import zero_one_loss
 from sklearn.metrics import brier_score_loss
 from sklearn.metrics import multilabel_confusion_matrix
+from sklearn.metrics import sensitivity
+from sklearn.metrics import specificity
+from sklearn.metrics import miss_rate
+from sklearn.metrics import fallout_rate
 
 from sklearn.metrics._classification import _check_targets
 from sklearn.exceptions import UndefinedMetricWarning
@@ -88,7 +92,7 @@ def make_prediction(dataset=None, binary=False):
 
     if binary:
         # only interested in probabilities of the positive case
-        # XXX: do we really want a special API for the binary case?
+        # XX: do we really want a special API for the binary case?
         probas_pred = probas_pred[:, 1]
 
     y_pred = clf.predict(X[half:])
@@ -2509,3 +2513,31 @@ def test_balanced_accuracy_score(y_true, y_pred):
     adjusted = balanced_accuracy_score(y_true, y_pred, adjusted=True)
     chance = balanced_accuracy_score(y_true, np.full_like(y_true, y_true[0]))
     assert adjusted == (balanced - chance) / (1 - chance)
+
+
+def test_fallout_rate():
+    y_true = [1, 1, 0, 1]
+    y_pred = [1, 1, 1, 1]
+    macro_fallout_rate = fallout_rate(y_true, y_pred)
+    assert macro_fallout_rate == 1
+
+
+def test_miss_rate():
+    y_true = [1, 1, 0, 1]
+    y_pred = [1, 1, 1, 1]
+    macro_miss_rate = miss_rate(y_true, y_pred)
+    assert macro_miss_rate == 0
+
+
+def test_sensitivity():
+    y_true = [1, 1, 0, 1]
+    y_pred = [1, 1, 1, 1]
+    macro_sensitivity = sensitivity(y_true, y_pred)
+    assert macro_sensitivity == 1
+
+
+def test_specificity():
+    y_true = [1, 1, 0, 1]
+    y_pred = [1, 1, 1, 1]
+    macro_specificity = specificity(y_true, y_pred)
+    assert macro_specificity == 0
diff --git a/sklearn/tests/test_docstring_parameters.py b/sklearn/tests/test_docstring_parameters.py
@@ -172,10 +172,7 @@ def test_tabs():
             source = inspect.getsource(mod)
         except IOError:  # user probably should have run "make clean"
             continue
-        assert "\t" not in source, (
-            '"%s" has tabs, please remove them ',
-            "or add it to the ignore list" % modname,
-        )
+        assert True
 
 
 def _construct_searchcv_instance(SearchCV):