def select(self, data_matrix, target=None):
if sparse.issparse(data_matrix):
data_matrix = SparseRandomProjection().fit_transform(data_matrix).toarray()
mf = pymf.SIVM(data_matrix.T, num_bases=self.n_instances)
mf.factorize()
basis = mf.W.T
selected_instances_ids = self._get_ids(data_matrix, basis)
return selected_instances_ids
def transform(self, data_matrix):
basis_data_matrix = self.matrix_factorizer.W
data_matrix_new = self.transformer.transform(data_matrix)
self.matrix_factorizer = pymf.SIVM(data_matrix_new.T, num_bases=self.complexity)
self.matrix_factorizer.W = basis_data_matrix
self.matrix_factorizer.factorize(compute_w=False)
if self.n_kmeans:
return self.kmeans.transform(self.matrix_factorizer.H.T)
else:
return self.matrix_factorizer.H.T
def fit(self, data_matrix):
n_rows, n_cols = data_matrix.shape
if n_rows <= n_cols:
n_components = n_rows
elif n_cols < 5000:
n_components = n_cols
else:
n_components = 'auto'
self.transformer = random_projection.SparseRandomProjection(n_components=n_components,
dense_output=True,
random_state=self.random_state)
data_matrix_new = self.transformer.fit_transform(data_matrix)
self.matrix_factorizer = pymf.SIVM(data_matrix_new.T, num_bases=self.complexity)
self.matrix_factorizer.factorize()
if self.n_kmeans:
self.kmeans = MiniBatchKMeans(n_clusters=self.n_kmeans)
self.kmeans.fit(self.matrix_factorizer.H.T)
def matrix_factorization(data_matrix, n=10):
mf = pymf.SIVM(data_matrix.T, num_bases=n)
mf.factorize()
return mf.W.T, mf.H.T
def select_layer(self, data_matrix):
mf = pymf.SIVM(data_matrix.T, num_bases=self.n_instances)
mf.factorize()
basis = mf.W.T
selected_instances_ids = self._get_ids(data_matrix, basis)
return selected_instances_ids