Abstract: We propose a saddlepoint approximation of the error probability of a binary hypothesis test between two i.i.d. distributions. The approximation is accurate ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results