Bayesian Nonparametric Methods for Learning Markov Switching Processes
Name
Fox-2010-Bayesian Nonparametr.pdf
Size
1.12 MB
Format
Adobe PDF
Checksum (MD5)
558dd58ac2b006b095b4c72b99d74c7b
Author(s)
Fox, Emily Beth
Willsky, Alan S.
Sudderth, Erik B.
Jordan, Michael I.
Date Issued
November 2010
Journal
IEEE Signal Processing Magazine
Publisher
Institute of Electrical and Electronics Engineers (IEEE)
Citation
Fox, Emily et al. “Bayesian Nonparametric Methods for Learning Markov Switching Processes.” IEEE Signal Processing Magazine (2010): n. pag. Web. 3 Feb. 2012. © 2011 Institute of Electrical and Electronics Engineers
Version
Final published version
Abstract
In this article, we explored a Bayesian nonparametric approach to learning Markov switching processes. This framework requires one to make fewer assumptions about the underlying dynamics, and thereby allows the data to drive the complexity of the inferred model. We began by examining a Bayesian nonparametric HMM, the sticky HDPHMM, that uses a hierarchical DP prior to regularize an unbounded mode space. We then considered extensions to Markov switching processes with richer, conditionally linear dynamics, including the HDP-AR-HMM and HDP-SLDS. We concluded by considering methods for transferring knowledge among multiple related time series. We argued that a featural representation is more appropriate than a rigid global clustering, as it encourages sharing of behaviors among objects while still allowing sequence-specific variability. In this context, the beta process provides an appealing alternative to the DP.
MIT Department
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
Terms of Use
Article is made available in accordance with the publisher's policy and may be subject to US copyright law. Please refer to the publisher's site for terms of use.
Persistent DSpace Link
DOI of Published Version
https://doi.org/10.1109/msp.2010.937999
