 |
Towards a Maximum Entropy Method for Estimating HMM Parameters
Walder, Christian J., Kootsookos, Peter J. and Lovell, Brian C. (2003). Towards a Maximum Entropy Method for Estimating HMM Parameters. In: Lovell, Brian C. and Maeder, Anthony J, Proceedings of the 2003 APRS Workshop on Digital Image Computing. Workshop on Digital Image Computing, Brisbane, (45-49). February 7, 2003.
|
|
| |
| Attached Files (Some files may be inaccessible until you login with your UQ eSpace credentials) |
| Name |
Description |
MIMEType |
Size |
Downloads |
n45.pdf
|
n45.pdf |
application/pdf |
74.93KB |
520 |
| Author(s) |
Walder, Christian J.
Kootsookos, Peter J.
Lovell, Brian C.
|
| Title of paper |
Towards a Maximum Entropy Method for Estimating HMM Parameters
|
| Conference name |
Workshop on Digital Image Computing
|
| Conference location |
Brisbane
|
| Conference dates |
February 7, 2003
|
| Proceedings title |
Proceedings of the 2003 APRS Workshop on Digital Image Computing
|
| Editor(s) |
Lovell, Brian C.
Maeder, Anthony J
|
| Place published |
Brisbane
|
| Publisher |
Australian Pattern Recognition Society
|
| Publication date |
2003
|
| Volume number |
1
|
| Issue number |
1
|
| ISBN |
0-9580255-2-5
|
| Start page |
45
|
| End page |
49
|
| Total pages |
5
|
| Language |
eng
|
| Abstract/Summary |
Training a Hidden Markov Model (HMM) to maximise the probability of a given sequence can result in over-fitting. That is, the model represents the training sequence well, but fails to generalise. In this paper, we present a possible solution to this problem, which is to maximise a linear combination of the likelihood of the training data, and the entropy of the model. We derive the necessary equations for gradient based maximisation of this combined term. The performance of the system is then evaluated in comparison with three other algorithms, on a classification task using synthetic data. The results indicate that the method is potentially useful. The main problem with the method is the computational intractability of the entropy calculation.
|
| Subjects |
280207 Pattern Recognition
E1
|
| Keyword(s) |
iris-research
hidden markov model
|
|
|
|
