Fitting sparse Markov models through a collapsed Gibbs sampler

Sparse Markov models (SMMs) provide a parsimonious representation for higher-order Markov models. We present a computationally efficient method for fitting SMMs using a collapsed Gibbs sampler, the GSDPMM. We prove the consistency of the GSDPMM in fitting SMMs. In simulations, the GSDPMM was found t...

Full description

Saved in:
Bibliographic Details
Published in:Computational statistics 2023-12, Vol.38 (4), p.1977-1994
Main Authors: Bennett, Iris, Martin, Donald E. K., Lahiri, Soumendra Nath
Format: Article
Language:English
Subjects:
Algorithms
Clustering
Economic Theory/Quantitative Economics/Mathematical Methods
Gene sequencing
Markov analysis
Markov chains
Mathematics and Statistics
Original Paper
Probability
Probability and Statistics in Computer Science
Probability Theory and Stochastic Processes
Samplers
Sparsity
Statistics
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Sparse Markov models (SMMs) provide a parsimonious representation for higher-order Markov models. We present a computationally efficient method for fitting SMMs using a collapsed Gibbs sampler, the GSDPMM. We prove the consistency of the GSDPMM in fitting SMMs. In simulations, the GSDPMM was found to perform as well or better than existing methods for fitting SMMs. We apply the GSDPMM method to fit SMMs to patterns of wind speeds and DNA sequences.
ISSN:0943-4062
1613-9658
DOI:10.1007/s00180-022-01310-8