期刊文献+

一种基于SSM的HMM训练算法 被引量:1

A HMM Training Algorithm Based on SSM
在线阅读 下载PDF
导出
摘要 在Baum Welch(BW)算法的基础上提出了一种基于态相关方法(State SpecificMethod:SSM)的隐马尔可夫模型(HiddenMarkovMode:HMM)参数估计算法(简称SBW算法).该算法在估计HMM不同状态的概率密度函数(proba bilitydensityfunction:PDF)的参数时使用了与状态有关的维数较低的特征集合.与传统的BW算法相比,新算法避免了直接估计高维的PDF参数.仿真实验表明,在训练数据量不足的情况下,采用SBW算法的误识率明显低于BW算法. In this paper, we derive an algorithm based on well-known BW algorithm for estimating the parameters of a hidden Markov model (we call it SBW algorithm). SBW algorithm relies on a low dimensional state-specific feature set rather than rely on a common high dimensional feature set as conventional BW algorithm, so avoiding directly estimating high-dimensional probability density functions in HMM training. Our computer simulation example shows that the performance of the new algorithm is superior (over) the conventional Baum-Welch algorithm.
出处 《武汉大学学报(理学版)》 CAS CSCD 北大核心 2003年第5期625-628,共4页 Journal of Wuhan University:Natural Science Edition
基金 湖北省教育厅重点项目基金资助项目(2002A02004)
关键词 态相关方法 隐马尔可夫模型 参数估计 Baum-Welch算法 SSM HMM训练算法 SBW算法 语音识别 hidden Markov model(HMM) Baum-Welch algorithm parameter estimation state-specific method sufficient statistics
  • 相关文献

参考文献2

二级参考文献14

  • 1王作英.基于段长分布的HMM语音识别模型.第二届全国汉字语音识别会议[M].庐山,1989..
  • 2-.智能机研究动态.第五届全国汉字识别、语音识别与合成系统及自然语言处理系统评测结果[M].,1994,4..
  • 3Wang Zuoying,Chin J Electron,1998年,7卷,1期
  • 4Wang Zuoying,Technical Report,1998年
  • 5Wang Zuoying,Proc the First Int Conference on Multimodal Interface(ICMI'96),1996年,70页
  • 6智能机研究动态,1994年
  • 7Wang Zuoying,Telecommunication Science,1993年,9卷,4期,31页
  • 8Young S J,HTK Version 1.5:User Reference and Programmer Manuals,1993年
  • 9王作英,第二届全国汉字语音识别会议,1989年
  • 10B. Mak, E. Bocchieri. Direct training of subspace distribution clustering hidden Markov model. IEEE Trans. Speech Audio Processing, 2001, 9(4): 378-387.

共引文献11

同被引文献12

  • 1[1]L.E.Baum,T.Petrie.Statistical inference for prob-abilistic function of finite state Markov chains [J].Annals of Math. Statistic,1996,(6):1554-1563.
  • 2[2]L.E.Baum,J.A.Egon.An inequality with applic-ations to statistical estimation for probabilisticfunctions of a Markov process and to a model for ecology [J].Bull.Amer. Meteorology , 1967,(2):360-363.
  • 3[3]L.Rabiner.A tutorial on hidden Markov modelsand selected applications in speech recognition[J].Proc. IEEE,1989,(2) :257-286.
  • 4[4]L.E.Baum,T.Petrie.A maximization technique occurring in the statistical analysis of probabilistic functions of Markov chains[J].Annals of Math. Statistic.1970,(1):164-171.
  • 5[5]L.E.Baum.An inequality and associated maximi-zation technique in statistical estimation for proba-bilistic functions of Markov processes[J].Inequali-ties,1970,(1):1-8.
  • 6[6]S.E.Leveson, L.Rabiner. An introduction to theapplication of the theory of probabilistic functions of Markov process to automatic speech recognition[J].Bell System Journal,1983,(4):1035-1047.
  • 7[7]L.R.Bahl,F.Jelinek,R,A,Mercer.A maximum likelihood approach to continuous speech recogni-tion[J]. IEEE Trans. Patten Analysis and Machine Intelligence,1983, (1) :179-190.
  • 8[9]M.J.F.Gales,Semi-tied covariance matrices for hidden Markov models[J].IEEE Trans.Speech Audio Processing,1999,(3):272-281.
  • 9[10]P.M.Bagggenstoss.Class-specific features inclassification[J].IEEE Trans.on Signal Process-ing,1999,(12):3428-3432.
  • 10[11]L.Rabiner,B.H.Juang.Fundamentals of SpeechRecognition[M].USA:Prentice-Hall Press,1993.

引证文献1

二级引证文献1

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部