MIDASVARResult._trailing_blocks#

MIDASVARResult._trailing_blocks()[source]#

The compressed-regressor coefficients, in design order.

Return type:

tuple[NDArray[double], ...]

Returns:

(midas_coefficients.T,), an (m, k) block whose rows match the compressed columns at the end of design.

Example

>>> import numpy as np
>>> rng = np.random.default_rng(0)
>>> x = rng.standard_normal((240, 1))
>>> last = x[2::3, 0] + 0.1 * rng.standard_normal(80)
>>> y = np.column_stack([last, rng.standard_normal(80)])
>>> res = MIDASVAR(y, x, order=1, period=3).fit()
>>> [block.shape for block in res._trailing_blocks()]
[(1, 2)]