def test_3d_error(self):
data = np.array(2)
with pytest.raises(ValueError):
stattools.lagmat2ds(data, 5)
data = np.zeros((100, 2, 2))
with pytest.raises(ValueError):
stattools.lagmat2ds(data, 5)
def test_lagmat2ds_pandas(self):
data = self.macro_df
lagmat = stattools.lagmat2ds(data, 2)
expected = self._prepare_expected(data.values, 2)
assert_array_equal(lagmat, expected)
lagmat = stattools.lagmat2ds(data.iloc[:, :2], 3, trim="both")
expected = self._prepare_expected(data.values[:, :2], 3)
expected = expected[3:]
assert_array_equal(lagmat, expected)
data = self.series
lagmat = stattools.lagmat2ds(data, 5)
expected = self._prepare_expected(data.values[:, None], 5)
assert_array_equal(lagmat, expected)
def test_lagmat2ds_numpy(self):
data = self.macro_df
npdata = data.values
lagmat = stattools.lagmat2ds(npdata, 2)
expected = self._prepare_expected(npdata, 2)
assert_array_equal(lagmat, expected)
lagmat = stattools.lagmat2ds(npdata[:, :2], 3)
expected = self._prepare_expected(npdata[:, :2], 3)
assert_array_equal(lagmat, expected)
npdata = self.series.values
lagmat = stattools.lagmat2ds(npdata, 5)
expected = self._prepare_expected(npdata[:, None], 5)
assert_array_equal(lagmat, expected)
def get_lagged_data(x, lag, addconst, verbose):
x = np.asarray(x)
# print x.shape[0]
# print x.shape
if x.shape[0] <= 3 * lag + int(addconst):
raise ValueError("Insufficient observations. Maximum allowable "
"lag is {0}".format(int((x.shape[0] - int(addconst)) /
3) - 1))
if verbose:
print('\nGranger Causality')
print('number of lags (no zero)', lag)
# create lagmat of both time series
dta = lagmat2ds(x, lag, trim='both', dropex=1)
#add constant
if addconst:
dtaown = add_constant(dta[:, 1:(lag + 1)], prepend=False, has_constant = 'add')
dtajoint = add_constant(dta[:, 1:], prepend=False, has_constant = 'add')
else:
dtaown = dta[:, 1:(lag+1)]
dtajoint = dta[:, 1:]
return dta, dtaown, dtajoint
def test_lagmat2ds_use_pandas(self):
data = self.macro_df
lagmat = stattools.lagmat2ds(data, 2, use_pandas=True)
expected = self._prepare_expected(data.values, 2)
cols = []
for c in data:
for lags in range(3):
if lags == 0:
cols.append(c)
else:
cols.append(c + ".L." + str(lags))
expected = pd.DataFrame(expected, index=data.index, columns=cols)
assert_frame_equal(lagmat, expected)
lagmat = stattools.lagmat2ds(data.iloc[:, :2],
3,
use_pandas=True,
trim="both")
expected = self._prepare_expected(data.values[:, :2], 3)
cols = []
for c in data.iloc[:, :2]:
for lags in range(4):
if lags == 0:
cols.append(c)
else:
cols.append(c + ".L." + str(lags))
expected = pd.DataFrame(expected, index=data.index, columns=cols)
expected = expected.iloc[3:]
assert_frame_equal(lagmat, expected)
data = self.series
lagmat = stattools.lagmat2ds(data, 5, use_pandas=True)
expected = self._prepare_expected(data.values[:, None], 5)
cols = []
c = data.name
for lags in range(6):
if lags == 0:
cols.append(c)
else:
cols.append(c + ".L." + str(lags))
expected = pd.DataFrame(expected, index=data.index, columns=cols)
assert_frame_equal(lagmat, expected)
