Move-to-front and inversion coding

Abstract

Motivated by the move-to-front (MTF) coder's (recency ranking) utilization of small size permutations which are restricted to the data source's alphabet size, we investigate compression of data files by using the canonical sorting permutations from the set of {0,...,n}, where n is the size of a data source, followed by inversion coding. We show that the inversion coding (ranking) technique introduced yields better compression gain than the recency ranking (MTF coder) for almost all the test data files. Furthermore, we investigate replacement of MTF coder in the block sorting schemes and show that when inversion coding is used after the BWT transformation, it yields better compression gains on average than the well-known block sorting techniques such as Bzip, Bks98 and Szip-b.