Application of the Zipf law to text compression
М. П. Бакулина · Journal of Applied and Industrial Mathematics · 2008
Under consideration is one of the important problems of information theory, the problem of the compression of data, in particular texts in natural languages, preserving the possibility of their unique restoration (decoding). A way of solution is proposed: to construct the codes that are based on the Zipf law. In distinction from the general methods, a construction of this kind uses the information about the statistical structure of the source of messages. The algorithms of two-pass and single-pass encoding schemes are adduced and the compression effectiveness is estimated.