fips.Posterior.with_data#
- Posterior.with_data(obs)[source]#
Return the posterior of the same problem with other observations.
The analytic method’s factorization depends on where the observations are and how uncertain they are, not on their values, so this costs a product with the factors instead of a new solve (a bootstrap, or synthetic observations of a known truth): one \(m \times m\) product in the whitened form, a solve with the model-data mismatch and \(O(n^2)\) in the information form.
- Parameters:
obs (
Vector) – New observations, with exactly the labels of the problem’s (in any order).- Returns:
The posterior of the problem with obs, its estimate computed.
- Return type:
- Raises:
ValueError – If obs does not have exactly the observations’ labels.
TypeError – If the method cannot take new observations without solving again (its solution has no
with_data).