def test_basic_outlier_filter(self):
np.random.seed(42)
x = np.random.normal(size=20)
x[0] *= 5
msk = basic_outlier_filter(x)
self.assertEqual(np.sum(~msk), 1)
self.assertAlmostEqual(x[~msk][0], 2.4835707650561636)
def filter_for_sparsity(data, c1=1e3, solver='ECOS'):
daily_sparsity = np.sum(data > 0.005 * np.max(data), axis=0)
filtered_signal = local_median_regression_with_seasonal(daily_sparsity,
c1=c1,
solver=solver)
mask = basic_outlier_filter(daily_sparsity - filtered_signal,
outlier_constant=5.)
return mask
def filter_for_sparsity(data, c1=1e3, solver="ECOS"):
daily_sparsity = np.sum(data > 0.005 * np.max(data), axis=0)
filtered_signal = l1_l2d2p365(daily_sparsity, c1=c1, solver=solver)
mask = basic_outlier_filter(daily_sparsity - filtered_signal, outlier_constant=5.0)
return mask