Next:
Introduction
Up:
A Time Varying Appearance
Previous:
Conclusion
Index
Variable length Markov model
Subsections
Introduction
Encoding transition probabilities
The storage of transitions
Towards a more effective storage of transitions
Training a VLMM
Definitions
The learning algorithm
General idea
The pruning of nodes
The algorithm
The estimation of observed probabilities
Laplace's law of succession
The maximum likelihood estimate
Lidstone's law of succession
The natural law of succession
Comparison of the probability distributions
The Kullback-Leibler divergence
The Matusita distance
Prediction using VLMM
VLMM results
Comparison of the Lidstone probability estimation with the maximum likelihood probability estimation
Comparison of trees built using the Matusita distance
Comparison of trees built using the Kullback-Leibler divergence
Comparison of trees built using the Matusita measure with trees built using the Kullback-Leibler divergence for large texts
Quantitative assessment of the prediction
Performance given a large training set
Performance given a small training set
Conclusion
franck 2006-10-01