Binary Hypothesis Testing using Structured Adaptive Networks
Ioannis N. M. Papadakis, Stelios C. A. Thomopoulos · 1992
The design of binary hypothesis tests in the absence of any statistical information, based only on a set of available observations is studied. A Structured Adaptive Network (SAN) configuration for the design of such tests based on several criteria of optimality, is presented and certain key asymptotic properties of these criteria are established.