Hidden Markov Models and Applications (Unsupervised and Semi-Supervised Learning)

✍ Scribed by Nizar Bouguila (editor), Wentao Fan (editor), Manar Amayri (editor)

• Publisher: Springer
• Year: 2022
• Tongue: English
• Leaves: 303
• Category: Library

✦ Synopsis

This book focuses on recent advances, approaches, theories, and applications related Hidden Markov Models (HMMs). In particular, the book presents recent inference frameworks and applications that consider HMMs. The authors discuss challenging problems that exist when considering HMMs for a specific task or application, such as estimation or selection, etc. The goal of this volume is to summarize the recent advances and modern approaches related to these problems. The book also reports advances on classic but difficult problems in HMMs such as inference and feature selection and describes real-world applications of HMMs from several domains. The book pertains to researchers and graduate students, who will gain a clear view of recent developments related to HMMs and their applications.

✦ Table of Contents

Preface
Contents
Contributors
A Roadmap to Hidden Markov Models and a Review of Its Application in Occupancy Estimation
1 Introduction
1.1 Objectives
1.2 Outline
2 Hidden Markov Models
2.1 Overview
2.2 Topologies
2.3 Gaussian Mixture Models and the Expectation–Maximization Algorithm
2.4 Baum Welch Algorithm
2.5 Viterbi Algorithm
2.6 Applications
3 Survey of the Employment of Hidden Markov Models in Occupancy Estimation
3.1 Limitations and Future Venues of Improvement
4 Conclusion
References
Bounded Asymmetric Gaussian Mixture-Based Hidden MarkovModels
1 Introduction
2 Bounded Asymmetric Gaussian Mixture Model
3 Hidden Markov Model
4 BAGMM Integration into the HMM Framework
4.1 Estimation of π and A
4.2 Estimation of Λ
4.3 Complete Algorithm
5 Experimental Results
5.1 Occupancy Estimation
5.1.1 Occupancy Detection Dataset
5.1.2 Occupancy Estimation Dataset
5.1.3 Experimental Results
5.2 Human Activity Recognition (HAR)
5.2.1 HAR Dataset
5.2.2 Preprocessing and Data Visualization
5.2.3 Methodology and Results
6 Conclusion
References
Using HMM to Model Neural Dynamics and Decode Useful Signals for Neuroprosthetic Control
1 Introduction
2 General Principles OF HMM
3 Neural Implementation of HMMs
4 HMMs of Parietal Cortex Activity During an Arm Movement Task
4.1 Functional Characterization of the Parietal Cortex Areas
4.2 Decoding of Task Epoch and Target Position
4.3 HMM: Parietal Cortex Functions and Information Decoding
5 Related Works
5.1 HMM Applications to Model Neurophysiological Data
5.2 Other Approaches to Model Neurophysiological Data
6 Conclusions
References
Fire Detection in Images with Discrete Hidden Markov Models
1 Introduction
2 Proposed System
2.1 Forward Algorithm
2.2 Baum–Welch Algorithm
2.3 HMM Framework
3 Experimental Results
3.1 Dataset
3.2 Evaluation Metrics
3.3 Result Discussions
3.3.1 Performance Evaluation
3.3.2 Comparative Analysis
4 Conclusion
References
Hidden Markov Models: Discrete Feature Selection in Activity Recognition
1 Introduction
2 Hidden Markov Models and Feature Selection
2.1 Hidden Markov Models
2.2 Feature Selection Methods
2.2.1 Filter-Based Techniques
2.2.2 Wrapper-Based Techniques
3 Experimental Setup and Results
3.1 Experimental Setup
3.2 Data
3.3 Evaluation Metrics
3.4 Results
3.4.1 Filter-Based Methods
3.4.2 Backward Elimination
4 Conclusion
References
Bayesian Inference of Hidden Markov Models Using DirichletMixtures
1 Introduction
2 The Learning Model
2.1 The Bayesian Model
2.2 Mixture Model
2.3 Hidden Markov Model
2.3.1 Prior Distributions
2.3.2 Complete Hierarchical Model
3 Markov Chain Monte Carlo Methodology
3.1 Gibbs Moves
3.2 Split and Combine Moves
3.3 Birth and Death Moves
4 Experiments
4.1 Human Activity Recognition
4.2 Speaker Recognition
5 Conclusion
Appendix
References
Online Learning of Inverted Beta-Liouville HMMs for Anomaly Detection in Crowd Scenes
1 Introduction
2 Related Work
3 Hidden Markov Models
3.1 Notations and Offline EM for HMMs
3.2 Online EM for HMMs
3.2.1 Sufficient Statistics for Parameter Estimation
3.2.2 Recurrence Relations
4 Inverted Beta-Liouville Mixture Model
4.1 Online Update for the Sufficient Statistics and Model Parameters
5 Experiments and Results
5.1 Anomaly Detection in a Crowd of Pedestrians
5.2 Anomaly Detection: Airport Security Line-Up
5.3 Abnormal Crowd Behavior: Escape Scene
6 Conclusion
References
A Novel Continuous Hidden Markov Model for Modeling Positive Sequential Data
1 Introduction
2 The HMM with Inverted Beta-Liouville Mixture Models
2.1 The Formulation of IBL-HMM
2.2 The Prior Distributions
3 Model Fitting by Variational Bayes
3.1 The Optimization of q(a a a a), q(π π π π), and q(c c c c)
3.2 The Optimization of q(λ λ λ λ), q(α α α α), and q(β β β β)
3.3 The Optimization of q(S,L)
4 Experimental Results
4.1 Data Sets and Experimental Settings
5 Conclusion
References
Multivariate Beta-Based Hidden Markov Models Applied to Human Activity Recognition
1 Introduction
2 Hidden Markov Model
3 HMM Parameters Estimation with Maximum Likelihood
4 HMM Parameters Estimation with Variational Inference
5 Experimental Results
6 Conclusion
References
Multivariate Beta-Based Hierarchical Dirichlet Process Hidden Markov Models in Medical Applications
1 Introduction
2 Model Specification
2.1 Multivariate Beta-Based Hidden Markov Model
2.2 Multivariate Beta-Based Hierarchical Dirichlet Process of Hidden Markov Model
3 Variational Learning
4 Experimental Results
4.1 First Individual, First Run of Activities
4.2 First Individual, Second Run of Activities
5 Conclusion
References
Shifted-Scaled Dirichlet-Based Hierarchical Dirichlet Process Hidden Markov Models with Variational Inference Learning
1 Introduction
2 Hidden Markov Models
3 Variational Learning
3.1 Shifted-Scaled Dirichlet-Based Hidden Markov Model
3.2 Shifted-Scaled-Based Hierarchical Dirichlet Process Hidden Markov Model
3.2.1 Update Q(Zv) and Q(θi)
3.2.2 Update Q(W), Q(ψ), and Q()
4 Experimental Results
4.1 Activity Recognition
4.2 Texture Clustering
5 Conclusion
Appendices
Appendix 1
Appendix 2
References
Index