Merge pull request #115 from dvro/datasets

[MRG] adding make_imbalance function

Author: dvro

examples/datasets/plot_make_imbalance.py

@@ -0,0 +1,56 @@
+"""
+===========================
+make_imbalance function
+===========================
+
+An illustration of the make_imbalance function
+
+"""
+
+print(__doc__)
+
+import numpy as np
+
+import matplotlib.pyplot as plt
+import seaborn as sns
+sns.set()
+
+# Define some color for the plotting
+almost_black = '#262626'
+palette = sns.color_palette()
+
+from sklearn.datasets import make_moons
+from imblearn.datasets import make_imbalance
+
+
+# Generate the dataset
+X, y = make_moons(n_samples=200, shuffle=True, noise=0.5, random_state=10)
+
+# Two subplots, unpack the axes array immediately
+f, axs = plt.subplots(2, 3)
+
+axs = [a for ax in axs for a in ax]
+
+axs[0].scatter(X[y == 0, 0], X[y == 0, 1], label="Class #0",
+            alpha=0.5, edgecolor=almost_black, facecolor=palette[0],
+            linewidth=0.15)
+axs[0].scatter(X[y == 1, 0], X[y == 1, 1], label="Class #1",
+           alpha=0.5, edgecolor=almost_black, facecolor=palette[2],
+           linewidth=0.15)
+axs[0].set_title('Original set')
+
+ratios = [0.9, 0.75, 0.5, 0.25, 0.1]
+for i, ratio in enumerate(ratios, start=1):
+    ax = axs[i]
+
+    X_, y_ = make_imbalance(X, y, ratio=ratio, min_c_=1)
+
+    ax.scatter(X_[y_ == 0, 0], X_[y_ == 0, 1], label="Class #0",
+                alpha=0.5, edgecolor=almost_black, facecolor=palette[0],
+                linewidth=0.15)
+    ax.scatter(X_[y_ == 1, 0], X_[y_ == 1, 1], label="Class #1",
+               alpha=0.5, edgecolor=almost_black, facecolor=palette[2],
+               linewidth=0.15)
+    ax.set_title('make_imbalance ratio ({})'.format(ratio))
+
+plt.show()

imblearn/datasets/__init__.py

@@ -0,0 +1,8 @@
+"""
+The :mod:`imblearn.datasets` provides methods to generate 
+imbalanced data.
+"""
+
+from .imbalance import make_imbalance
+
+__all__ = ['make_imbalance']

imblearn/datasets/imbalance.py

@@ -0,0 +1,74 @@
+"""Transform a dataset into an imbalanced dataset."""
+
+import numpy as np
+
+from collections import Counter
+
+from sklearn.utils import check_X_y
+from sklearn.utils import check_random_state
+
+def make_imbalance(X, y, ratio, min_c_=None, random_state=None):
+    """Turns a dataset into an imbalanced dataset at specific ratio.
+    A simple toy dataset to visualize clustering and classification
+    algorithms.
+
+    Parameters
+    ----------
+    X : ndarray, shape (n_samples, n_features)
+        Matrix containing the data to be imbalanced.
+
+    y : ndarray, shape (n_samples, )
+        Corresponding label for each sample in X.
+
+    ratio : float, 
+        The desired ratio given by the number of samples in 
+        the minority class over the the number of samples in 
+        the majority class.
+
+    min_c_ : str or int, optional (default=None)
+        The identifier of the class to be the minority class.
+        If None, min_c_ is set to be the current minority class.
+
+    random_state : int, RandomState instance or None, optional (default=None)
+        If int, random_state is the seed used by the random number generator;
+        If RandomState instance, random_state is the random number generator;
+        If None, the random number generator is the RandomState instance used
+        by np.random.
+
+    Returns
+    -------
+    X_resampled : ndarray, shape (n_samples_new, n_features)
+        The array containing the imbalanced data.
+
+    y_resampled : ndarray, shape (n_samples_new)
+        The corresponding label of `X_resampled`
+    """
+    if ratio <= 0.0 or ratio >= 1.0:
+        raise ValueError('ratio value must be such that 0.0 < ratio < 1.0')
+
+    X, y = check_X_y(X, y)
+
+    random_state = check_random_state(random_state)
+
+    stats_c_ = Counter(y)
+
+    if min_c_ is None:
+        min_c_ = min(stats_c_, key=stats_c_.get)
+
+    n_min_samples = int(np.count_nonzero(y != min_c_) * ratio)
+    if n_min_samples > stats_c_[min_c_]:
+        raise ValueError('Current imbalance ratio of data is lower than desired ratio!')
+    if n_min_samples == 0:
+        raise ValueError('Not enough samples for desired ratio!')
+
+    mask = y == min_c_
+
+    idx_maj = np.where(~mask)[0]
+    idx_min = np.where(mask)[0]
+    idx_min = random_state.choice(idx_min, size=n_min_samples, replace=False)
+    idx = np.concatenate((idx_min, idx_maj), axis=0)
+
+    X_resampled, y_resampled = X[idx,:], y[idx]
+
+    return X_resampled, y_resampled
+

imblearn/datasets/tests/__init__.py

@@ -0,0 +1,108 @@
+"""Test the module easy ensemble."""
+from __future__ import print_function
+
+import numpy as np
+from numpy.testing import assert_raises
+from numpy.testing import assert_equal
+
+from collections import Counter
+
+from imblearn.datasets import make_imbalance
+
+# Generate a global dataset to use
+X = np.random.random((1000, 2))
+Y = np.zeros(1000)
+Y[500:] = 1
+
+def test_make_imbalance_bad_ratio():
+    """Test either if an error is raised with bad ratio
+    argument"""
+    min_c_ = 1
+
+    # Define a zero ratio
+    ratio = 0.0
+    assert_raises(ValueError, make_imbalance, X, Y, ratio, min_c_)
+
+    # Define a negative ratio
+    ratio = -2.0
+    assert_raises(ValueError, make_imbalance, X, Y, ratio, min_c_)
+
+    # Define a ratio greater than 1
+    ratio = 2.0
+    assert_raises(ValueError, make_imbalance, X, Y, ratio, min_c_)
+
+    # Define ratio as a list which is not supported
+    ratio = [.5, .5]
+    assert_raises(ValueError, make_imbalance, X, Y, ratio, min_c_)
+
+
+def test_make_imbalance_invalid_ratio():
+    """Test either if error is raised with higher ratio
+    than current ratio."""
+
+    y_ = np.zeros((X.shape[0], ))
+    y_[0] = 1
+
+    ratio = 0.5
+    assert_raises(ValueError, make_imbalance, X, y_, ratio)
+
+def test_make_imbalance_single_class():
+    """Test either if an error when there is a single class"""
+    y_ = np.zeros((X.shape[0], ))
+    ratio = 0.5
+    assert_raises(ValueError, make_imbalance, X, y_, ratio)
+
+def test_make_imbalance_1():
+    """Test make_imbalance"""
+    X_, y_ = make_imbalance(X, Y, ratio=0.5, min_c_=1)
+    counter = Counter(y_)
+    assert_equal(counter[0], 500)
+    assert_equal(counter[1], 250)
+    assert(np.all([X_i in X for X_i in X_]))
+
+def test_make_imbalance_2():
+    """Test make_imbalance"""
+    X_, y_ = make_imbalance(X, Y, ratio=0.25, min_c_=1)
+    counter = Counter(y_)
+    assert_equal(counter[0], 500)
+    assert_equal(counter[1], 125)
+    assert(np.all([X_i in X for X_i in X_]))
+
+def test_make_imbalance_3():
+    """Test make_imbalance"""
+    X_, y_ = make_imbalance(X, Y, ratio=0.1, min_c_=1)
+    counter = Counter(y_)
+    assert_equal(counter[0], 500)
+    assert_equal(counter[1], 50)
+    assert(np.all([X_i in X for X_i in X_]))
+
+def test_make_imbalance_4():
+    """Test make_imbalance"""
+    X_, y_ = make_imbalance(X, Y, ratio=0.01, min_c_=1)
+    counter = Counter(y_)
+    assert_equal(counter[0], 500)
+    assert_equal(counter[1], 5)
+    assert(np.all([X_i in X for X_i in X_]))
+
+def test_make_imbalance_5():
+    """Test make_imbalance"""
+    X_, y_ = make_imbalance(X, Y, ratio=0.01, min_c_=0)
+    counter = Counter(y_)
+    assert_equal(counter[1], 500)
+    assert_equal(counter[0], 5)
+    assert(np.all([X_i in X for X_i in X_]))
+
+def test_make_imbalance_multiclass():
+    """Test make_imbalance with multiclass data"""
+    # Make y to be multiclass
+    y_ = np.zeros(1000)
+    y_[100:500] = 1
+    y_[500:] = 2
+
+    # Resample the data
+    X_, y_ = make_imbalance(X, y_, ratio=0.1, min_c_=0)
+    counter = Counter(y_)
+    assert_equal(counter[0], 90)
+    assert_equal(counter[1], 400)
+    assert_equal(counter[2], 500)
+    assert(np.all([X_i in X for X_i in X_]))