Generalized binary Bayesian receiver
Abstract
A description is presented of an approach which generalizes the Bayesian binary receiver in the sense that it can also be applied to situations in which the available information about the binary source is not sufficient to define a prior probability distribution. The approach makes use of an axiomatic structure of upper and lower probabilities. The obtained system transforms prior knowledge into an intervalvalued expected loss over the space of decision. The approach supplies, thus, a systematic way of transforming an unreliable prior knowledge into a family of posterior probability distributions, which combined with the loss function provides an intervalvalue support over the action space.
 Publication:

ICC '81; International Conference on Communications, Volume 2
 Pub Date:
 1981
 Bibcode:
 1981icc.....2...30W
 Keywords:

 Bayes Theorem;
 Binary Data;
 Digital Systems;
 Probability Distribution Functions;
 Signal Reception;
 Communication Theory;
 Invariance;
 Linear Transformations;
 Probability Theory;
 Transition Probabilities;
 Communications and Radar