Cadeias estocásticas com memória de alcance variável

Abstract

Stochastic chains with memory of variable length, consists of analyzing a portion of the past, called context, to predict the next symbol. Recently, they have been used to model up scientific data in areas as different as biology, linguistics and music. This work seeks to define this family of stochastic chains, present the maximum likelihood estimation of transition probabilities given a context tree, present the algorithm context, the BIC method and compare these two methods in software R.