A special type of graphical model. The observed variable, Y, has a distribution that depends on the value of the corresponding X-variable. The idea is that the successive values of X obey the Markov property, with the value of each depending only on the value of its predecessor.