二阶隐Markov模型的一种简化算法及参数估计

摘要

An abridged algorithm of second-order Hidden Markov Models(HMM2) was made. New algorithm was similar to the well-known Baum-Welch algorithm for estimating the parameters of a Hidden Markov Models(HMM), which allowed the observation PDF of each state to be defined and estimated u-sing a different feature set. Based on this method the corresponding recongnition method and parameter estimation are presented. This method was based on sufficient statistics and resuited in no theoretical loss of performance.%针对现有的二阶隐Markov模型(second-order HMM:HMM2)算法给出了一种新的简化算法.新算法是一种与Baum-Welch算法相似的算法,允许每一状态的观测概率密度函数(probability density function:PDF)可以用不同的特征集来定义和估计.基于此方法给出了相应的识别方法和参数估计.这种方法是基于充分统计量的,并且从理论上来说不会造成性能的损失.