def _sprandn(m, n, density=0.01, format="coo", dtype=None, random_state=None):
# Helper function for testing.
if random_state is None:
random_state = np.random
elif isinstance(random_state, (int, np.integer)):
random_state = np.random.RandomState(random_state)
data_rvs = random_state.randn
return construct.random(m, n, density, format, dtype, random_state, data_rvs)
def _sprandn(m, n, density=0.01, format="coo", dtype=None, random_state=None):
# Helper function for testing.
if random_state is None:
random_state = np.random
elif isinstance(random_state, (int, np.integer)):
random_state = np.random.RandomState(random_state)
data_rvs = random_state.randn
return construct.random(m, n, density, format, dtype, random_state,
data_rvs)
def test_random_sparse_matrix_returns_correct_number_of_non_zero_elements(
self):
# A 10 x 10 matrix, with density of 12.65%, should have 13 nonzero elements.
# 10 x 10 x 0.1265 = 12.65, which should be rounded up to 13, not 12.
sparse_matrix = construct.random(10, 10, density=0.1265)
assert_equal(sparse_matrix.count_nonzero(), 13)
def test_random_accept_str_dtype(self):
# anything that np.dtype can convert to a dtype should be accepted
# for the dtype
construct.random(10, 10, dtype='d')
def _sprandn(m, n, density=0.01, format="coo", dtype=None, random_state=None):
# Helper function for testing.
random_state = check_random_state(random_state)
data_rvs = random_state.standard_normal
return construct.random(m, n, density, format, dtype, random_state,
data_rvs)
def test_random_accept_str_dtype(self):
# anything that np.dtype can convert to a dtype should be accepted
# for the dtype
a = construct.random(10, 10, dtype='d')
# X_2 = np.where(X > 0.6 , 1, 0) #X = X_1 + X_2 # maintain 0 issue index = np.where(X==0) check = index[1] pair = np.where(check%2==0, check+1, check-1) newindex = (index[0], pair) print newindex X[newindex] = 0 B = sparse.random(p, 1, density=0.1) # B is random sparse vector B = B.A B = np.around(B, decimals=2) Y_1 = X.dot(B) Y = np.around(Y_1, decimals=2) print X.shape print B.shape print Y.shape # Y = np.where(Y_1 > np.median(Y_1, axis=0), 2,1) # Clustering # clf = KMeans(n_clusters = k) # s = clf.fit(Y_1)
