.. _DefBLAdjust:

Definition of Term: Bayes linear adjustment
===========================================

In the context of the :ref:`Bayes linear<DefBayesLinear>` approach,
suppose that we have specified prior means, variances and covariances
for a collection of uncertain quantities. If the values of some of the
quantities become known, then we reassess our judgements about the
remaining quantities. As the Bayes linear approach is
:ref:`Bayesian<DefBayesian>`, we modify our prior beliefs in light of
the data, which (under the Bayes linear approach) yields the *adjusted
expectation* and *adjusted variance* for each quantity and the *adjusted
covariance* for every pair.

See the procedure page on calculation of adjusted expectation and
variance (:ref:`ProcBLAdjust<ProcBLAdjust>`) for details on
calculating the adjusted expectations, variances and covariances, and
the discussion page on theoretical aspects of Bayes linear
(:ref:`DiscBayesLinearTheory<DiscBayesLinearTheory>`) for further
discussion of the theory behind Bayes linear.