Error detection in a class of decimal codes
S. Fratini · IEEE Transactions on Information Theory · 1989
Consideration is given to the problem of multiple parity checks for an n-digit number. Each check relates to specific digits, where the probability p of a particular digit being in error is fixed. The author proves that it is best, in the sense of minimizing the probability of an undetected error, to add either one or n parity checks, depending on the value of n and p.>