International Conference on Intelligent Control and Information Processing , 2010-08-13

Title : ( Hybrid of Chaos Optimization and Baum-Welch Algorithms for HMM Training in Continuous Speech Recognition )

Authors: somayeh cheshomi , saeed rahati , Mohammad Reza Akbarzadeh Totonchi ,

Citation: BibTeX | EndNote

Abstract

In this paper a new optimization algorithm based on Chaos Optimization algorithm(COA) combined with traditional Baum Welch (BW) method is presented for training Hidden Markov Model (HMM) for Continues speech recognition. The BW algorithm easily trapped in local optimum . which might deteriorate the speech recognition rate, while an important character of COA is global search .so we can get a globally optimal solution or at least sub-optimal solution. In this paper Chaos optimization algorithm was applied to the optimization of the initial value of HMM parameters in Baum-Welch algorithm. Experimental results showed that using Chaos Optimization algorithm for HMM training (Chaos-HMM training) has a better performance than using other heuristic algorithms such as PSOBW and GAPSOBW

Keywords

, Hybrid of Chaos Optimization and Baum, Welch Algorithms for HMM Training in Continuous Speech Recognition
برای دانلود از شناسه و رمز عبور پرتال پویا استفاده کنید.

@inproceedings{paperid:1019479,
author = {Somayeh Cheshomi and Saeed Rahati and Akbarzadeh Totonchi, Mohammad Reza},
title = {Hybrid of Chaos Optimization and Baum-Welch Algorithms for HMM Training in Continuous Speech Recognition},
booktitle = {International Conference on Intelligent Control and Information Processing},
year = {2010},
keywords = {Hybrid of Chaos Optimization and Baum-Welch Algorithms for HMM Training in Continuous Speech Recognition},
}

[Download]

%0 Conference Proceedings
%T Hybrid of Chaos Optimization and Baum-Welch Algorithms for HMM Training in Continuous Speech Recognition
%A Somayeh Cheshomi
%A Saeed Rahati
%A Akbarzadeh Totonchi, Mohammad Reza
%J International Conference on Intelligent Control and Information Processing
%D 2010

[Download]