Computer Science · Artificial Intelligence
The mixture transition distribution model for highorder Markov chains and non-Gaussian time series
Simon Tavare, Adrian E. Raftery, Simon Tavare, Adrian E. Rta:ttei'y
See on the mapAbstract
The mixture transition distribution (MTD) model was introduced by Raftery (1985) as a parsimonious model for high-order Markov chains. It is flexible, can a wide range of patterns, can be physically motivated, data well, and appears to be a discrete-valued analogue for the class of autoregressive series models. However, estimation has presented difficulties because the space is highly nonconvex, being defined by a large number of nonlinear constraints. we propose an computational algorithm for maximum likelihood is on a way of reducing the of also stnlcttlred verSl011S of Away apli>liE!d to a sequence 1