Title :
Exponential-type error probabilities for multiterminal hypothesis testing
Author :
Han, Te Sun ; Kobayashi, Kingo
Author_Institution :
Dept. of Inf. Syst., Senshu Univ., Kawasaki, Japan
fDate :
1/1/1989 12:00:00 AM
Abstract :
Multiterminal hypothesis testing is considered, subject to the exponential-type constraint αn⩽exp(-nr) on the error probability of the first kind. The problem is to determine the minimum β*n of the error probability of the second kind under the given constraint at limited rates R1 and R2 for observing the respective pairs of variables. Good lower bounds on the power exponent for β*n are presented by invoking basic properties of r-divergent sequences. In particular, the one-bit and the zero-rate compression cases are considered. The power exponent for the former and a lower bound for the latter are established
Keywords :
error statistics; information theory; probability; error probability; exponential-type constraint; information theory; lower bounds; multiterminal hypothesis testing; one-bit compression; power exponent; r-divergent sequences; zero-rate compression; Decoding; Error probability; Information systems; Information theory; Mathematics; Performance evaluation; Statistical analysis; System testing;
Journal_Title :
Information Theory, IEEE Transactions on