Title :
The convolution inequality for entropy powers
Author :
Blachman, Nelson M.
fDate :
4/1/1965 12:00:00 AM
Abstract :
The entropy power of a band-limited random process is the power of white Gaussian noise having the same entropy rate. Shannon´s convolution inequality for entropy power states that the entropy power of the sum of two independent random processes is at least the sum of their entropy powers. This paper presents an improved version of Stam´s proof of this inequality, which is obtained by mathematical induction from the one-dimensional case.
Keywords :
Convolution; Entropy functions; Bandwidth; Convolution; Covariance matrix; Entropy; Gaussian noise; Gaussian processes; Random processes; Random variables; Statistics; Upper bound;
Journal_Title :
Information Theory, IEEE Transactions on
DOI :
10.1109/TIT.1965.1053768