Minimum-Entropy Decision Analysis
, by Alfonso Gonzalez-Serrano
The concept of minimum-entropy decision analysis is reviewed. Maximum
likelihood is defined as a process which optimizes error probability,
thus minimizing entropy. Some bounds on error probability are derived
based on the Tchebyscheff inequality and the union bound. The concepts
are illustrated with the additive white Gaussian noise channel. An
example for a picking algorithm is discussed, where performance is
optimized by introducing memory in the system.
STANFORD