Title :
Bayes Risk Error is a Bregman Divergence
Author :
Varshney, Kush R.
Author_Institution :
Bus. Analytics & Math. Sci. Dept., IBM T. J. Watson Res. Center, Yorktown Heights, NY, USA
Abstract :
In previous work reported in these Transactions, we proposed a new distortion measure for the quantization of prior probabilities that are used in the threshold of likelihood ratio test detection: Bayes risk error. In this correspondence, we show that the Bayes risk error is a member of the class of Bregman divergences and discuss the implications of this fact.
Keywords :
Bayes methods; probability; quantisation (signal); signal detection; Bayes risk error; Bregman divergence; distortion measurement; likelihood ratio test detection threshold; probability quantization; Bayesian methods; Context; Distortion measurement; Humans; Quantization; Signal detection; Testing; Bayesian hypothesis testing; Bregman divergence; quantization; signal detection;
Journal_Title :
Signal Processing, IEEE Transactions on
DOI :
10.1109/TSP.2011.2159500
