def test_corner():
# fails because lag < 1
with pytest.raises(ValueError):
diff(x=x, lag=0)
with pytest.raises(ValueError):
diff_inv(x=x, lag=0)
# fails because differences < 1
with pytest.raises(ValueError):
diff(x=x, differences=0)
with pytest.raises(ValueError):
diff_inv(x=x, differences=0)
# Passing in xi with the incorrect shape to a 2-d array
with pytest.raises(IndexError):
diff_inv(x=np.array([[1, 1], [1, 1]]), xi=np.array([[1]]))
def test_diff():
# test vector for lag = (1, 2), diff = (1, 2)
assert_array_equal(diff(x, lag=1, differences=1), np.ones(4))
assert_array_equal(diff(x, lag=1, differences=2), np.zeros(3))
assert_array_equal(diff(x, lag=2, differences=1), np.ones(3) * 2)
assert_array_equal(diff(x, lag=2, differences=2), np.zeros(1))
# test matrix for lag = (1, 2), diff = (1, 2)
assert_array_equal(diff(m, lag=1, differences=1),
np.array([[-5, -5, -2], [7, -15, 12]]))
assert_array_equal(diff(m, lag=1, differences=2),
np.array([[12, -10, 14]]))
assert_array_equal(diff(m, lag=2, differences=1), np.array([[2, -20, 10]]))
assert diff(m, lag=2, differences=2).shape[0] == 0
def test_adf_ols():
# Test the _ols function of the ADF test
x = np.array([1, -1, 0, 2, -1, -2, 3])
k = 2
y = diff(x)
assert_array_equal(y, [-2, 1, 2, -3, -1, 5])
z = ADFTest._embed(y, k).T
res = ADFTest._ols(x, y, z, k)
# Assert on the params of the OLS. The comparisons are those obtained
# from the R function.
expected = np.array([1.0522, -3.1825, -0.1609, 1.4690])
assert np.allclose(res.params, expected, rtol=0.001)
# Now assert on the standard error
stat = ADFTest._ols_std_error(res)
assert np.allclose(stat, -100.2895) # derived from R code
def test_corner():
# fails because lag < 1
with pytest.raises(ValueError):
diff(x=x, lag=0)
In this example, we demonstrate pyramid's array differencing, and how it's used in conjunction with the ``d`` term to lag a time series. .. raw:: html <br/> """ print(__doc__) # Author: Taylor Smith <*****@*****.**> from pmdarima.utils import array # Build an array and show first order differencing results x = array.c(10, 4, 2, 9, 34) lag_1 = array.diff(x, lag=1, differences=1) # The result will be the same as: x[1:] - x[:-1] print(lag_1) # [-6., -2., 7., 25.] # Note that lag and differences are not the same! If we crank diff up by one, # it performs the same differencing as above TWICE. Lag, therefore, controls # the number of steps backward the ts looks when it differences, and the # `differences` parameter controls how many times to repeat. print(array.diff(x, lag=1, differences=2)) # [4., 9., 18.] # Conversely, when we set lag to 2, the array looks two steps back for its # differencing operation (only one). print(array.diff(x, lag=2, differences=1)) # [-8., 5., 32.] # The lag parameter is controlled by `m`, which is the seasonal periodicity of
