NormalInverseWishartPrior._minnesota#

NormalInverseWishartPrior._minnesota()[source]#

The equivalent MinnesotaPrior, cross-equation weight pinned at one.

Every moment this prior reports is computed by that object, so the two classes cannot drift apart in their variance rules; this class adds only the pinned weight, the absence of a sum-of-coefficients field, and its own label. Built on each call, which is cheap, since the prior is a frozen dataclass of four scalars.

Return type:

MinnesotaPrior

Returns:

A MinnesotaPrior with cross_equation=1.0 and sum_of_coefficients=None, the other fields copied.

Example

>>> NormalInverseWishartPrior(tightness=0.1)._minnesota()._label()
'minnesota(l1=0.1, l2=1, l3=1, l4=100)'