Next: Loss functions for approximation
Up: Bayesian decision theory
Previous: Bayesian decision theory
  Contents
In Bayesian decision theory
a set of possible actions is considered,
together with a function
describing the loss suffered
in situation if appears and action is selected
[16,128,185,201].
The loss averaged over test data , ,
and possible states of Nature
is known as expected risk,
where
denotes the expectation under the joint predictive density
=
and
|
(44) |
The aim is to find an optimal action
|
(45) |
Joerg_Lemm
2001-01-21