VARMA.lag_order_selection#

VARMA.lag_order_selection(max_lags=None)[source]#

Unavailable: VARMA orders are not nested least-squares candidates.

The autoregressive scan the base class runs compares pure lag structures on one design family; a VARMA’s (p, q) trades the two polynomials off against each other, and scoring that honestly means fitting each candidate in full. Fit the specifications under consideration and rank them with information_criteria – the comparison surface exists precisely for this.

Parameters:

max_lags (int | None) – Ignored.

Raises:

SpecificationError – Always.

Return type:

NoReturn

Example

>>> import numpy as np
>>> y = np.random.default_rng(0).standard_normal((200, 2))
>>> VARMA(y, order=(1, 1)).lag_order_selection()
Traceback (most recent call last):
cultivars.exceptions.SpecificationError: lag-order selection over a single ...