A measure for the length of probabilistic dependence

A very useful model to describe a random process with memory is the Markov model. The MLSE receiver based on a finite order Markov model is studied in Kong and Shwedyk (1995). It is found that the order of the Markov model, i.e., the length of probabilistic dependence, is the crucial parameter that determines the receiver complexity. An information-theoretic measure for the length of probabilistic dependence is proposed.