基于扩展的HMM观察序列概率计算  

The Probability Calculation of Observation Sequence based on the Extended HMM

作　　者：菅小艳[1] 

机构地区：[1]太原师范学院计算机系,太原030012

出　　处：《电脑开发与应用》2011年第2期3-4,共2页Computer Development & Applications

摘　　要：HMM在自然语言处理领域占有很重要的地位。通常可以解决三大问题,主要针对第一个问题:给定HMM,计算观察序列的概率。提出了一种HMM的扩展模型,利用该扩展模型,计算观察序列概率的时间复杂度为T,大大节约了开支。但该模型也存在一定的不足之处,也是以后的研究方向。HMM plays an important role in the field of natural language processing.Normally,three major issues can be resolved,this paper aimed at the first question： for a given HMM,calculate the probability of observation.We proposed the expansion of an HMM model,Using the extended model to calculate the probability of observation sequence,time complexity is T and much cost sere saved.However,there are also some shortcomings,but also the future research directions.

关 键 词：HMM 前向算法 隐藏状态 观察状态 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]