Higher order adaptive arithmetic coding using the prediction by partial match algorithm

Arithmetic coding completely bypasses the traditional coding paradigm that replaces an input symbol by a specific code. Arithmetic coding represents a stream of input symbols by a single codeword. The modelling paradigm is a key aspect in effective data compression. We present an adaptive higher order arithmetic coding scheme using the prediction by partial match (PPM) algorithm as well as a framework for the comparison of the different compression techniques. Simulation study conducted with this framework shows that compressive effectiveness and implementing complexity increases as the order increases but above a certain order there is no gain in compression with increasing costs.