NormalInverseWishartPrior.exogenous

NormalInverseWishartPrior.exogenous: float

Looseness of the deterministic and exogenous block \(\lambda_4\).

The intercept, any trend, and any exogenous regressor get standard deviation \(\lambda_1 \lambda_4 s_i\), four orders of magnitude looser than an own lag at the default. Large but finite by design: a marginal likelihood exists only under a proper prior, and this is the field an improper “flat” intercept prior would have to go through. Must be positive.