Hidden Markov Models (HMMs)

For more info about our services contact : help@bestpfe.com

Table of contents

1 Introduction 
2 Representations, models and offline algorithms 
2.1 Audio signal representation
2.2 Bregman divergences
2.2.1 Definition and examples
2.2.2 Duality
2.3 Clustering with Bregman divergences
2.3.1 Bregman centroids
2.3.2 K-means
2.3.3 Bregman divergences and exponential families
2.3.4 Mixture models and the EM algorithm
2.4 Hidden Markov Models (HMMs)
2.4.1 Model
2.4.2 Inference
2.4.3 EM algorithm
2.5 Hidden Semi-Markov Models (HSMMs)
2.5.1 Model
2.5.2 Inference
2.5.3 EM algorithm
2.6 Summary and discussion
3 Online algorithms 
3.1 Online EM
3.1.1 Online EM for HMMs
3.1.2 Online EM for HSMMs
3.2 Non-probabilistic models
3.2.1 Online algorithm
3.3 Incremental EM
3.3.1 Incremental EM for HMMs
3.3.2 Semi-Markov extension
3.4 Including prior knowledge with Bayesian priors
3.5 Experiments on synthetic data
3.6 Summary and discussion
4 Audio segmentation experiments 
4.1 Offline segmentation results
4.2 Online EM for HMMs and HSMMs
4.3 Online vs incremental EM for HMMs
4.4 Segmentation of acoustic scenes
4.5 Summary and discussion