def fit(self, X, y=None, **fit_params):
"""Train the self-organizing map.
Parameters
----------
X : array-like or sparse matrix, shape=(n_samples, n_features)
Training instances to cluster.
y : Ignored
"""
# Check and normalize input data
X = minmax_scale(check_array(X, dtype=np.float32))
# Initialize Somoclu object
if not hasattr(self, 'algorithm_'):
# Set number of columns and rows from number of clusters
if self.n_clusters is not None:
self.n_columns_ = self.n_rows_ = int(self.n_clusters *
(np.sqrt(len(X)) - 2) + 2)
else:
self.n_columns_, self.n_rows_ = self.n_columns, self.n_rows
# Create object
self.algorithm_ = Somoclu(n_columns=self.n_columns_,
n_rows=self.n_rows_,
initialcodebook=self.initialcodebook,
kerneltype=self.kerneltype,
maptype=self.maptype,
gridtype=self.gridtype,
compactsupport=self.compactsupport,
neighborhood=self.neighborhood,
std_coeff=self.std_coeff,
initialization=self.initialization,
data=None,
verbose=self.verbose)
# Fit Somoclu
self.algorithm_.train(data=X, **fit_params)
# Grid labels
grid_labels = [
tuple(grid_label) for grid_label in self.algorithm_.bmus
]
# Generate labels mapping
labels_mapping = self._generate_labels_mapping(grid_labels)
# Generate cluster labels
self.labels_ = np.array(
[labels_mapping[grid_label] for grid_label in grid_labels])
# Generate labels neighbors
self.neighbors_ = self._generate_neighbors(grid_labels, labels_mapping)
return self
def test_deterministic_codebook(self):
n_rows, n_columns = 2, 2
codebook = np.zeros((2*2, 2), dtype=np.float32)
data = np.array([[0.1, 0.2], [0.3, 0.4]], dtype=np.float32)
som = Somoclu(n_columns, n_rows, initialcodebook=codebook,
compactsupport=False)
som.train(data)
correct_codebook = np.array([[[ 0.2 , 0.30000001],
[ 0.10359724, 0.20359723]],
[[ 0.29640275, 0.39640275],
[ 0.2 , 0.30000001]]], dtype=np.float32)
self.assertTrue(sum(codebook.reshape((n_rows*n_columns*2)) -
correct_codebook.reshape((n_rows*n_columns*2))) < 10e-8)
def fit(self, X, y=None, **fit_params):
"""Train the self-organizing map.
Parameters
----------
X : array-like or sparse matrix, shape=(n_samples, n_features)
Training instances to cluster.
y : Ignored
"""
# Check and normalize input data
X = minmax_scale(check_array(X, dtype=np.float32))
# Check random_state
self.random_state_ = check_random_state(self.random_state)
# Initialize codebook
if self.initialcodebook is None:
if self.random_state is None:
initialcodebook = None
initialization = 'random'
else:
codebook_size = self.n_columns * self.n_rows * X.shape[1]
initialcodebook = self.random_state_.random_sample(
codebook_size).astype(np.float32)
initialization = None
elif self.initialcodebook == 'pca':
initialcodebook = None
initialization = 'random'
else:
initialcodebook = self.initialcodebook
initialization = None
# Create Somoclu object
self.algorithm_ = Somoclu(
n_columns=self.n_columns,
n_rows=self.n_rows,
initialcodebook=initialcodebook,
kerneltype=self.kerneltype,
maptype=self.maptype,
gridtype=self.gridtype,
compactsupport=self.compactsupport,
neighborhood=self.neighborhood,
std_coeff=self.std_coeff,
initialization=initialization,
data=None,
verbose=self.verbose,
)
# Fit Somoclu
self.algorithm_.train(data=X, **fit_params)
# Grid labels
grid_labels = [
tuple(grid_label) for grid_label in self.algorithm_.bmus
]
# Generate labels mapping
self.labels_mapping_ = self._generate_labels_mapping(grid_labels)
# Generate cluster labels
self.labels_ = np.array(
[self.labels_mapping_[grid_label] for grid_label in grid_labels])
# Generate labels neighbors
self.neighbors_ = self._generate_neighbors(
np.unique(grid_labels, axis=0), self.labels_mapping_)
return self