Extended binomial moments of a linear code and the undetected error probability (Q2487096)

This paper introduces and studies the extended binomial moments of a linear code \(C\). Equalities relating the extended binomial moments of \(C\) and \(C^\perp\) have been derived. These lead to strict lower and upper bounds on the remaining nonzero extended binomial moments, and to a refinement of known inequalities between binomial moments, further leading to a simple algorithm for computing the code distance of \(C^\perp\) via the extended binomial moments of \(C\). The author further studies the probability of undetected error of \(C\) when the code is used for error detection on a symmetric channel. As a consequence, he derives a lower bound on the undetected error probability involving the basic code parameters and the number of minimum-weight codewords, and proves that the bound is attained by the near MDS and MMD codes.