On a class of arithmetic codes and a decoding algorithm (Corresp.)

Multiple-error-correcting arithmetic codes which are nonlinear are constructed by residue encoding. A simple algorithm is given for correcting multiple errors which makes use of continued fractions.