Merge pull request #49 from rasbt/tfcluster

Tfcluster

ci/.travis_test.sh

@@ -6,12 +6,14 @@ if [ "$TENSORFLOW" == "true" ]; then
     if [[ "$COVERAGE" == "true" ]]; then
         nosetests -s -v mlxtend/tf_classifier --nologcapture --with-coverage
         nosetests -s -v mlxtend/tf_regressor --nologcapture --with-coverage
+        nosetests -s -v mlxtend/tf_cluster --nologcapture --with-coverage
     else
         nosetests -s -v mlxtend/tf_classifier --nologcapture
         nosetests -s -v mlxtend/tf_regressor --nologcapture
+        nosetests -s -v mlxtend/tf_cluster --nologcapture
 else
     if [[ "$COVERAGE" == "true" ]]; then
-        nosetests -s -v --with-coverage --exclude-dir=mlxtend/tf_classifier --exclude-dir=mlxtend/tf_regressor --exclude-dir=mlxtend/data --exclude-dir=mlxtend/general_plotting
+        nosetests -s -v --with-coverage --exclude-dir=mlxtend/tf_classifier --exclude-dir=mlxtend/tf_regressor --exclude-dir=mlxtend/tf_cluster --exclude-dir=mlxtend/data --exclude-dir=mlxtend/general_plotting
     else
         nosetests -s -v --exclude-dir=mlxtend/tf_classifier --exclude-dir=mlxtend/data --exclude-dir=mlxtend/general_plotting
     fi

docs/mkdocs.yml

@@ -57,6 +57,8 @@ pages:
     - user_guide/feature_extraction/RBFKernelPCA.md
   - cluster:
     - user_guide/cluster/Kmeans.md
+  - tf_cluster:
+    - user_guide/tf_cluster/TfKmeans.md
   - evaluate:
     - user_guide/evaluate/confusion_matrix.md
     - user_guide/evaluate/plot_decision_regions.md
@@ -107,7 +109,8 @@ pages:
   - api_subpackages/mlxtend.evaluate.md
   - api_subpackages/mlxtend.feature_selection.md
   - api_subpackages/mlxtend.feature_extraction.md
-  - api_subpackages/mlxtend.cluster.md  
+  - api_subpackages/mlxtend.cluster.md
+  - api_subpackages/mlxtend.tf_cluster.md
   - api_subpackages/mlxtend.file_io.md
   - api_subpackages/mlxtend.general_plotting.md
   - api_subpackages/mlxtend.preprocessing.md

docs/sources/CHANGELOG.md

@@ -8,6 +8,7 @@
 
 - New TensorFlow estimator for Linear Regression ([`tf_regressor.TfLinearRegression`](./user_guide/tf_regressor/TfLinearRegression.md))
 - New k-means clustering estimator ([`cluster.Kmeans`](./user_guide/cluster/Kmeans.md))
+- New TensorFlow k-means clustering estimator ([`tf_cluster.Kmeans`](./user_guide/tf_cluster/TfKmeans.md))
 
 ##### Changes
 

docs/sources/USER_GUIDE_INDEX.md

@@ -34,6 +34,9 @@
 ## `cluster`
 - [`Kmeans`](user_guide/cluster/Kmeans.md) (new in 0.4.1dev)
 
+## `tf_cluster`
+- [`TfKmeans`](user_guide/tf_cluster/TfKmeans.md) (new in 0.4.1dev)
+
 ## `evaluate`
 - [`confusion_matrix`](user_guide/evaluate/confusion_matrix.md)
 - [`plot_decision_regions`](user_guide/evaluate/plot_decision_regions.md)

mlxtend/cluster/base.py

@@ -79,9 +79,10 @@ def _shuffle(self, arrays):
         r = np.random.permutation(len(arrays[0]))
         return [ary[r] for ary in arrays]
 
-    def _print_progress(self, iteration, cost=None, time_interval=10):
+    def _print_progress(self, iteration, n_iter,
+                        cost=None, time_interval=10):
         if self.print_progress > 0:
-            s = '\rIteration: %d/%d' % (iteration, self.n_iter)
+            s = '\rIteration: %d/%d' % (iteration, n_iter)
             if cost:
                 s += ' | Cost %.2f' % cost
             if self.print_progress > 1:
@@ -96,7 +97,7 @@ def _print_progress(self, iteration, cost=None, time_interval=10):
                         self.eta_str_ = '00:00:00'
                     if not iteration % time_interval:
                         eta_sec = ((ela_sec / float(iteration)) *
-                                   self.n_iter - ela_sec)
+                                   n_iter - ela_sec)
                         self.eta_str_ = self._to_hhmmss(eta_sec)
                     s += ' | ETA: %s' % self.eta_str_
             stderr.write(s)

mlxtend/cluster/kmeans.py

@@ -9,7 +9,7 @@
 from .base import _BaseCluster
 import numpy as np
 from time import time
-from scipy.spatial.distance import euclidean
+# from scipy.spatial.distance import euclidean
 
 
 class Kmeans(_BaseCluster):
@@ -86,12 +86,17 @@ def _fit(self, X):
                 self.centroids_ = new_centroids
 
             self.iterations_ += 1
+            if self.print_progress:
+                self._print_progress(iteration=self.iterations_,
+                                     n_iter=self.max_iter)
 
         return self
 
     def _get_cluster_idx(self, X, centroids):
         for sample_idx, sample in enumerate(X):
-            dist = [euclidean(sample, c) for c in self.centroids_]
+            # dist = [euclidean(sample, c) for c in self.centroids_]
+            dist = np.sqrt(np.sum(np.square(sample - self.centroids_), axis=1))
+
             yield np.argmin(dist)
 
     def _predict(self, X):

mlxtend/cluster/tests/test_kmeans.py

@@ -25,7 +25,7 @@ def test_nonfitted():
                   X)
 
 
-def test_three_blobs_1():
+def test_three_blobs_multi():
     km = Kmeans(k=3,
                 max_iter=50,
                 random_seed=1,
@@ -34,7 +34,18 @@ def test_three_blobs_1():
     assert (y_pred == y).all()
 
 
-def test_three_blobs_2():
+def test_three_blobs_1sample():
+    km = Kmeans(k=3,
+                max_iter=50,
+                random_seed=1,
+                print_progress=0)
+    sample = X[1, :].reshape(1, 2)
+
+    y_pred = km.fit(X).predict(sample)
+    assert y_pred[0] == y[1]
+
+
+def test_three_blobs_centroids():
     km = Kmeans(k=3,
                 max_iter=50,
                 random_seed=1,

mlxtend/tf_cluster/__init__.py

@@ -0,0 +1,9 @@
+# Sebastian Raschka 2014-2016
+# mlxtend Machine Learning Library Extensions
+# Author: Sebastian Raschka <sebastianraschka.com>
+#
+# License: BSD 3 clause
+
+from .tf_kmeans import TfKmeans
+
+__all__ = ["TfKmeans"]

mlxtend/tf_cluster/tests/test_tf_base.py

@@ -0,0 +1,71 @@
+# Sebastian Raschka 2014-2016
+# mlxtend Machine Learning Library Extensions
+# Author: Sebastian Raschka <sebastianraschka.com>
+#
+# License: BSD 3 clause
+
+from mlxtend.tf_cluster.tf_base import _TfBaseCluster
+import numpy as np
+from mlxtend.utils import assert_raises
+
+
+def test_init():
+    cl = _TfBaseCluster(print_progress=0, random_seed=1)
+
+
+def test_check_array_1():
+    X = np.array([1, 2, 3])
+    cl = _TfBaseCluster(print_progress=0, random_seed=1)
+    assert_raises(ValueError,
+                  'X must be a 2D array. Try X[:, numpy.newaxis]',
+                  cl._check_array,
+                  X)
+
+
+def test_check_array_2():
+    X = list([[1], [2], [3]])
+    cl = _TfBaseCluster(print_progress=0, random_seed=1)
+
+    assert_raises(ValueError,
+                  'X must be a numpy array',
+                  cl._check_array,
+                  X)
+
+
+def test_check_array_3():
+    X = np.array([[1], [2], [3]])
+    cl = _TfBaseCluster(print_progress=0, random_seed=1)
+    cl._check_array(X)
+
+
+def test_fit():
+    X = np.array([[1], [2], [3]])
+    tfr = _TfBaseCluster(print_progress=0, random_seed=1)
+    tfr.fit(X)
+
+
+def test_predict_1():
+    X = np.array([[1], [2], [3]])
+    cl = _TfBaseCluster(print_progress=0, random_seed=1)
+
+    assert_raises(AttributeError,
+                  'Model is not fitted, yet.',
+                  cl.predict,
+                  X)
+
+
+def test_predict_2():
+    X = np.array([[1], [2], [3]])
+    cl = _TfBaseCluster(print_progress=0, random_seed=1)
+
+    cl.fit(X)
+    cl.predict(X)
+
+
+def test_shuffle():
+    X = np.array([[1], [2], [3]])
+    y = np.array([1, 2, 3])
+    cl = _TfBaseCluster(print_progress=0, random_seed=1)
+    X_sh, y_sh = cl._shuffle(arrays=[X, np.array(y)])
+    np.testing.assert_equal(X_sh, np.array([[1], [3], [2]]))
+    np.testing.assert_equal(y_sh, np.array([1, 3, 2]))

mlxtend/tf_cluster/tests/test_tf_kmeans.py

@@ -0,0 +1,59 @@
+# Sebastian Raschka 2014-2016
+# mlxtend Machine Learning Library Extensions
+# Author: Sebastian Raschka <sebastianraschka.com>
+#
+# License: BSD 3 clause
+
+from mlxtend.data import three_blobs_data
+from mlxtend.tf_cluster import TfKmeans
+from mlxtend.utils import assert_raises
+import numpy as np
+
+
+X, y = three_blobs_data()
+
+
+def test_nonfitted():
+    km = TfKmeans(k=3,
+                  max_iter=50,
+                  random_seed=1,
+                  print_progress=0)
+
+    assert_raises(AttributeError,
+                  'Model is not fitted, yet.',
+                  km.predict,
+                  X)
+
+
+def test_three_blobs_multi():
+    km = TfKmeans(k=3,
+                  max_iter=50,
+                  random_seed=1,
+                  print_progress=0)
+    y_pred = km.fit(X).predict(X)
+    assert (y_pred == y).all()
+
+
+def test_three_blobs_1sample():
+    km = TfKmeans(k=3,
+                  max_iter=50,
+                  random_seed=1,
+                  print_progress=0)
+    sample = X[1, :].reshape(1, 2)
+
+    y_pred = km.fit(X).predict(sample)
+    assert y_pred[0] == y[1]
+
+
+def test_three_blobs_centroids():
+    km = TfKmeans(k=3,
+                  max_iter=50,
+                  random_seed=1,
+                  print_progress=0)
+
+    centroids = np.array([[-1.5947298, 2.92236966],
+                          [2.06521743, 0.96137409],
+                          [0.9329651, 4.35420713]])
+
+    km.fit(X)
+    np.testing.assert_almost_equal(centroids, km.centroids_, decimal=2)

mlxtend/tf_cluster/tf_base.py

@@ -0,0 +1,120 @@
+# Sebastian Raschka 2014-2016
+# mlxtend Machine Learning Library Extensions
+#
+# Base Clusteer (Clutering Parent Class)
+# Author: Sebastian Raschka <sebastianraschka.com>
+#
+# License: BSD 3 clause
+
+import numpy as np
+from sys import stderr
+from time import time
+import tensorflow as tf
+
+
+class _TfBaseCluster(object):
+
+    """Parent Class Base Cluster
+
+    A base class that is implemented by
+    clustering child classes.
+
+    """
+    def __init__(self, print_progress=0, random_seed=None, dtype=None):
+        self.print_progress = print_progress
+        self.random_seed = random_seed
+        if dtype is None:
+            self.dtype = tf.float32
+        else:
+            self.dtype = dtype
+        self._is_fitted = False
+
+    def fit(self, X):
+        """Learn cluster centroids from training data.
+
+        Parameters
+        ----------
+        X : {array-like, sparse matrix}, shape = [n_samples, n_features]
+            Training vectors, where n_samples is the number of samples and
+            n_features is the number of features.
+
+        Returns
+        -------
+        self : object
+
+        """
+        self._is_fitted = False
+        self._check_array(X=X)
+        if self.random_seed is not None:
+            np.random.seed(self.random_seed)
+        self._fit(X=X)
+        self._is_fitted = True
+        return self
+
+    def _fit(self, X):
+        # Implemented in child class
+        pass
+
+    def predict(self, X):
+        """Predict cluster labels of X.
+
+        Parameters
+        ----------
+        X : {array-like, sparse matrix}, shape = [n_samples, n_features]
+            Training vectors, where n_samples is the number of samples and
+            n_features is the number of features.
+
+        Returns
+        ----------
+        cluster_labels : array-like, shape = [n_samples]
+          Predicted cluster labels.
+
+        """
+        self._check_array(X=X)
+        if not self._is_fitted:
+            raise AttributeError('Model is not fitted, yet.')
+        return self._predict(X)
+
+    def _predict(self, X):
+        # Implemented in child class
+        pass
+
+    def _shuffle(self, arrays):
+        """Shuffle arrays in unison."""
+        r = np.random.permutation(len(arrays[0]))
+        return [ary[r] for ary in arrays]
+
+    def _print_progress(self, iteration, n_iter,
+                        cost=None, time_interval=10):
+        if self.print_progress > 0:
+            s = '\rIteration: %d/%d' % (iteration, n_iter)
+            if cost:
+                s += ' | Cost %.2f' % cost
+            if self.print_progress > 1:
+                if not hasattr(self, 'ela_str_'):
+                    self.ela_str_ = '00:00:00'
+                if not iteration % time_interval:
+                    ela_sec = time() - self.init_time_
+                    self.ela_str_ = self._to_hhmmss(ela_sec)
+                s += ' | Elapsed: %s' % self.ela_str_
+                if self.print_progress > 2:
+                    if not hasattr(self, 'eta_str_'):
+                        self.eta_str_ = '00:00:00'
+                    if not iteration % time_interval:
+                        eta_sec = ((ela_sec / float(iteration)) *
+                                   n_iter - ela_sec)
+                        self.eta_str_ = self._to_hhmmss(eta_sec)
+                    s += ' | ETA: %s' % self.eta_str_
+            stderr.write(s)
+            stderr.flush()
+
+    def _to_hhmmss(self, sec):
+        m, s = divmod(sec, 60)
+        h, m = divmod(m, 60)
+        return "%d:%02d:%02d" % (h, m, s)
+
+    def _check_array(self, X):
+        if isinstance(X, list):
+            raise ValueError('X must be a numpy array')
+        if not len(X.shape) == 2:
+            raise ValueError('X must be a 2D array. Try X[:, numpy.newaxis]')