ARFIMAResult.has_long_memory#

property ARFIMAResult.has_long_memory: bool[source]#

Whether d is far enough from zero to imply hyperbolic decay.

A threshold on the point estimate at 1e-3 in either direction, not a test; white noise typically returns a d of a few hundredths and so reads as long memory here.

Example

>>> import numpy as np
>>> rng = np.random.default_rng(0)
>>> res = ARFIMA(rng.standard_normal(400), order=(0, 0), trend="n").fit()
>>> res.has_long_memory == (abs(res.d) > 1e-3)
True