... Lemm1
Email: lemm@uni-muenster.de, WWW: http://pauli.uni-muenster.de/${}^\sim$lemm/
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... `field'.2
We may also remark that for example statistical field theories, which encompass quantum mechanics and quantum field theory in their Euclidean formulation, are technically similar to a nonparametric Bayesian approach [249,103,127].
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
....3
Besides $\phi $ also the hyperparameters discussed in Chapter 5 belong to the hidden variables $h$.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... prior.4
The variable $\beta $ is, together with the regression function $\phi(x)$, part of the hidden variables $\tilde \phi$ = $(\phi,\beta)$ which determine the likelihood and thus not a hyperparameter. Technically, however, $\beta $ can be treated similar to a regularization factor, being the reason that it is included in this chapter on hyperparameters.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.