MarkovSwitchingVARResult._conditional_mean#

static MarkovSwitchingVARResult._conditional_mean(model, stack, deterministic)[source]#

One regime’s one-step conditional means over the effective sample.

Lays the regime’s blocks out against the model’s [deterministic | lags] design and multiplies through.

Parameters:

  • model (_MarkovSwitchingVectorAutoRegressionModel[MarkovSwitchingVARResult]) – The model, read for its design matrix.

  • stack (NDArray[double]) – The regime’s (p, k, k) lag coefficients.

  • deterministic (NDArray[double]) – The regime’s (d, k) deterministic block.

Return type:

NDArray[double]

Returns:

A (nobs, k) array of conditional means as if the regime held at every date.

Example

>>> import numpy as np
>>> rng = np.random.default_rng(0)
>>> s = np.repeat([0, 1, 0, 1], 100)
>>> y = np.array([[-1.0, -0.5], [1.5, 1.0]])[s] + 0.4 * rng.standard_normal((400, 2))
>>> model = MarkovSwitchingVAR(y, order=1, n_regimes=2)
>>> res = model.fit(seed=0)
>>> view = res.regimes[1]
>>> mean = MarkovSwitchingVARResult._conditional_mean(
...     model, view.coefficients, view.deterministic
... )
>>> hand = view.deterministic[0] + y[:-1] @ view.coefficients[0].T
>>> mean.shape, bool(np.allclose(mean, hand))
((399, 2), True)