The Math Forum

Search All of the Math Forum:

Views expressed in these public forums are not endorsed by NCTM or The Math Forum.

Math Forum » Discussions » Software » comp.soft-sys.matlab

Notice: We are no longer accepting new posts, but the forums will continue to be readable.

Topic: Speech Recognition via MFCC(feature extraction) and HMM classfication
Replies: 7   Last Post: Mar 31, 2014 6:38 AM

Advanced Search

Back to Topic List Back to Topic List Jump to Tree View Jump to Tree View   Messages: [ Previous | Next ]

Posts: 19
Registered: 9/18/10
Speech Recognition via MFCC(feature extraction) and HMM classfication
Posted: Sep 18, 2010 8:04 AM
  Click to see the message monospaced in plain text Plain Text   Click to reply to this topic Reply


currently i doing my thesis(follows as the title) i have done the MFCC as the FE. now i doing the training part and testing phase.
for training i using the Baum-Welch algorithm for training and viterbi for recognition.

1)my question is my log-likelihood from my training set is in positive value. i am not sure whether it acceptable o not? most of the value are in positive but in some papers the value was in negative.

2) i have got some viterbi path. not sure how to make use of it? help me in that too

3) Actually i am confuse how to do the the recognition part? can anyone willing explain to me. how to do a effective recognition?

My matlab code:

%%%%%%%%Training of HMM (Baum-Welch algorithm)%%%%%%
[LL, prior1, transmat1, mu1, Sigma1, mixmat1] = ...
mhmm_em(data1, pi, transmat, mu, Sigma, mixmat, 'max_iter', 5);
mixmat1 = max(mixmat1, 1.0e-5);

%%%% Calculate the Log Likelihood %%%%%%
[loglik, errors] = mhmm_logprob(data1, prior1, transmat1, mu1, Sigma1, mixmat1);

%%%%%%%%%%%%%% Viterbi Alogorithm %%%%%%
B = mixgauss_prob(data1, mu1, Sigma1, mixmat1);
path1 = viterbi_path(prior1, transmat1, B);

i really hope someone can help me in this.. i would really appreciate it and once done i will share my code for your reference and future students as well..thanks ...this is my email if wana contact me.


Point your RSS reader here for a feed of the latest messages in this topic.

[Privacy Policy] [Terms of Use]

© The Math Forum at NCTM 1994-2018. All Rights Reserved.