0.14 release (#463)

docs/sources/CHANGELOG.md

@@ -7,7 +7,7 @@ The CHANGELOG for the current development version is available at
 
 ---
 
-### Version 0.14.0dev (TBD)
+### Version 0.14.0 (11-09-2018)
 
 ##### Downloads
 

mlxtend/__init__.py

@@ -4,4 +4,4 @@
 #
 # License: BSD 3 clause
 
-__version__ = '0.14.0dev'
+__version__ = '0.14.0'

mlxtend/evaluate/__init__.py

@@ -24,7 +24,7 @@
 from .holdout import PredefinedHoldoutSplit
 from .f_test import ftest
 from .f_test import combined_ftest_5x2cv
-
+from .proportion_difference import proportion_difference
 
 
 __all__ = ["scoring", "confusion_matrix",
@@ -36,4 +36,5 @@
            "paired_ttest_kfold_cv", "paired_ttest_5x2cv",
            "feature_importance_permutation",
            "RandomHoldoutSplit", "PredefinedHoldoutSplit",
-           "ftest", "combined_ftest_5x2cv"]
+           "ftest", "combined_ftest_5x2cv",
+           "proportion_difference"]

mlxtend/evaluate/proportion_difference.py

@@ -0,0 +1,51 @@
+# Sebastian Raschka 2014-2018
+# mlxtend Machine Learning Library Extensions
+#
+# Author: Sebastian Raschka <sebastianraschka.com>
+#
+# License: BSD 3 clause
+
+import numpy as np
+import scipy.stats
+
+
+def proportion_difference(proportion_1, proportion_2, n_1, n_2=None):
+    """
+    Computes the test statistic and p-value for a difference of
+    proportions test.
+
+    Parameters
+    -----------
+    proportion_1 : float
+        The first proportion
+    proportion_2 : float
+        The second proportion
+    n_1 : int
+        The sample size of the first test sample
+    n_2 : int or None (default=None)
+        The sample size of the second test sample.
+        If `None`, `n_1`=`n_2`.
+
+    Returns
+    -----------
+
+    z, p : float or None, float
+        Returns the z-score and the p-value
+
+
+    Examples
+    -----------
+    For usage examples, please see
+    http://rasbt.github.io/mlxtend/user_guide/evaluate/proportion_difference/
+
+    """
+    if n_2 is None:
+        n_2 = n_1
+
+    var_1 = proportion_1*(1. - proportion_1) / n_1
+    var_2 = proportion_2*(1. - proportion_2) / n_2
+
+    z = (proportion_1 - proportion_2) / np.sqrt(var_1 + var_2)
+    p = scipy.stats.norm.cdf(z)
+
+    return z, p

mlxtend/evaluate/tests/test_proportion_difference.py

@@ -0,0 +1,53 @@
+# Sebastian Raschka 2014-2018
+# mlxtend Machine Learning Library Extensions
+# Author: Sebastian Raschka <sebastianraschka.com>
+#
+# License: BSD 3 clause
+
+from mlxtend.evaluate import proportion_difference
+import numpy as np
+
+
+def test_on_dataset():
+    y_true = np.array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+                       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+                       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+                       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+                       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+                       0, 0, 0, 0, 0])
+
+    ym1 = np.array([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0,
+                    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+                    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+                    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+                    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+                    0, 0])
+
+    ym2 = np.array([1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+                    1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+                    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+                    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+                    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+                    0, 0])
+
+    ym3 = np.array([1, 1, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+                    1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+                    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+                    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+                    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+                    1, 1])
+
+    acc_1 = np.sum(y_true == ym1) / y_true.shape[0]
+    acc_2 = np.sum(y_true == ym2) / y_true.shape[0]
+    acc_3 = np.sum(y_true == ym3) / y_true.shape[0]
+    assert round(acc_1, 2) == 0.84
+    assert round(acc_2, 2) == 0.92
+    assert round(acc_3, 2) == 0.92
+
+    z, p_value = proportion_difference(acc_1, acc_2, n_1=y_true.shape[0])
+    assert round(z, 3) == -1.754
+    assert round(p_value, 3) == 0.040
+
+    z, p_value = proportion_difference(acc_2, acc_3, n_1=y_true.shape[0])
+    assert round(z, 3) == 0.0
+    assert round(p_value, 3) == 0.5