Learning hidden Markov models from aggregate observations

DOI10.1016/J.AUTOMATICA.2021.110100MaRDI QIDQ2071943zbMATH OpenOpenAlexFDO

Authors Rahul Singh, Qinsheng Zhang, Yongxin Chen

Publication date 31 January 2022

Published in Automatica (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2011.11236

zbMATH Keywords

expectation-maximization algorithmhidden Markov modelsparameter learningaggregate observations

Mathematics Subject Classification ID

Markov processes: estimation; hidden Markov models (62M05) Mathematical geography and demography (91D20)

Abstract: In this paper, we propose an algorithm for estimating the parameters of a time-homogeneous hidden Markov model from aggregate observations. This problem arises when only the population level counts of the number of individuals at each time step are available, from which one seeks to learn the individual hidden Markov model. Our algorithm is built upon expectation-maximization and the recently proposed aggregate inference algorithm, the Sinkhorn belief propagation. As compared with existing methods such as expectation-maximization with non-linear belief propagation, our algorithm exhibits convergence guarantees. Moreover, our learning framework naturally reduces to the standard Baum-Welch learning algorithm when observations corresponding to a single individual are recorded. We further extend our learning algorithm to handle HMMs with continuous observations. The efficacy of our algorithm is demonstrated on a variety of datasets.

Recommendations

• A spectral algorithm for learning hidden Markov models
• Tensor train spectral method for learning of hidden Markov models (HMM)
• Maximum-likelihood estimation for hidden Markov models
• Online Learning with Hidden Markov Models
• Simulated annealing for maximum a posteriori parameter estimation of hidden Markov models

Cites work

• scientific article; zbMATH DE number 3945181 (Why is no real title available?)
• scientific article; zbMATH DE number 4080514 (Why is no real title available?)
• scientific article; zbMATH DE number 48812 (Why is no real title available?)
• scientific article; zbMATH DE number 3492022 (Why is no real title available?)
• scientific article; zbMATH DE number 3567782 (Why is no real title available?)
• scientific article; zbMATH DE number 1222290 (Why is no real title available?)
• scientific article; zbMATH DE number 3318337 (Why is no real title available?)
• A Maximization Technique Occurring in the Statistical Analysis of Probabilistic Functions of Markov Chains
• An inequality with applications to statistical estimation for probabilistic functions of Markov processes and to a model for ecology
• Estimating discrete Markov models from various incomplete data schemes
• Estimation in Markov Models from Aggregate Data
• Estimation of Time-Varying Markov Processes with Aggregate Data
• Graphical models, exponential families, and variational inference
• Inference in hidden Markov models.
• Machine learning. A probabilistic perspective
• Multi-Marginal Optimal Transport and Probabilistic Graphical Models
• On the relation between optimal transport and Schrödinger bridges: a stochastic control viewpoint
• Pattern recognition and machine learning.
• Probabilistic graphical models.
• Sample-based population observers
• Stochastic control liaisons. Richard Sinkhorn meets Gaspard Monge on a Schrödinger bridge

Cited in

(4)

• A spectral algorithm for learning hidden Markov models
• A New Method for Learning Imprecise Hidden Markov Models
• scientific article; zbMATH DE number 5198952 (Why is no real title available?)
• Learning hidden Markov models with unknown number of states

Describes a project that uses

Uses Software

• PMTK
• PRMLT

This page was built for publication: Learning hidden Markov models from aggregate observations

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2071943)