Abstract
The Continuous-Time Hidden Markov Model (CT-HMM) is an attractive modeling tool for mHealth data that takes the form of events occurring at irregularly-distributed continuous time points. However, the lack of an efficient parameter learning algorithm for CT-HMM has prevented its widespread use, necessitating the use of very small models or unrealistic constraints on the state transitions. In this paper, we describe recent advances in the development of efficient EM-based learning methods for CT-HMM models. We first review the structure of the learning problem, demonstrating that it consists of two challenges: (1) the estimation of posterior state probabilities and (2) the computation of end-state conditioned expectations. The first challenge can be addressed by reformulating the estimation problem in terms of an equivalent discrete time-inhomogeneous hidden Markov model. The second challenge is addressed by exploiting computational methods traditionally used for continuous-time Markov chains and adapting them to the CT-HMM domain. We describe three computational approaches and analyze the tradeoffs between them. We evaluate the resulting parameter learning methods in simulation and demonstrate the use of models with more than 100 states to analyze disease progression using glaucoma and Alzheimer’s Disease datasets.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
- 2.
- 3.
Data were obtained from the ADNI database (adni.loni.usc.edu). The ADNI was launched in 2003 as a public-private partnership, led by Principal Investigator Michael W. Weiner, MD. The primary goal of ADNI has been to test whether serial magnetic resonance imaging (MRI), positron emission tomography (PET), other biological markers, and clinical and neuropsychological assessment can be combined to measure the progression of mild cognitive impairment (MCI) and early Alzheimer’s disease (AD). For up-to-date information, see http://www.adni-info.org.
- 4.
References
Al-Mohy, A.H., Higham, N.J.: Computing the Fréchet derivative of the matrix exponential, with an application to condition number estimation. SIAM Journal on Matrix Analysis and Applications 30(4), 1639–1657 (2009)
Al-Mohy, A.H., Higham, N.J.: Computing the action of the matrix exponential, with an application to exponential integrators. SIAM Journal on Scientific Computing 33(2), 488–511 (2011)
Bartolomeo, N., Trerotoli, P., Serio, G.: Progression of liver cirrhosis to HCC: An application of hidden Markov model. BMC Medical Research Methodology 11(38) (2011)
Bauer, F.L., Fike, C.T.: Norms and exclusion theorems. Numerische Mathematik 2(1), 137–141 (1960)
Bindel, D., Goodman, J.: Principles of scientific computing (2009)
Bladt, M., Sørensen, M.: Statistical inference for discretely observed Markov jump processes. J. R. Statist. Soc. B 39(3), 395–410 (2005)
Cox, D.R., Miller, H.D.: The Theory of Stochastic Processes. Chapman and Hall, London (1965)
Fagan, A.M., Head, D., Shah, A.R., et. al: Decreased CSF A beta 42 correlates with brain atrophy in cognitively normal elderly. Ann Neurol. 65(2), 176–183 (2009)
Golub, G.H., Van Loan, C.F.: Matrix computations, vol. 3. JHU Press (2012)
Higham, N.: Functions of Matrices: Theory and Computation. SIAM Press (2008)
Hobolth, A., Jensen, J.L.: Statistical inference in evolutionary models of DNA sequences via the EM algorithm. Statistical Applications in Genetics and Molecular Biology 4(1) (2005)
Hobolth, A., Jensen, J.L.: Summary statistics for endpoint-conditioned continuous-time Markov chains. Journal of Applied Probability 48(4), 911–924 (2011)
Jackson, C.H.: Multi-state models for panel data: The MSM package for R. Journal of Statistical Software 38(8) (2011)
Jensen, A.: Markoff chains as an aid in the study of Markoff processes. Skand. Aktuarietidskr 36, 87–91 (1953)
Kingman, S.: Glaucoma is second leading cause of blindness globally. Bulletin of the World Health Organization 82(11) (2004)
Leiva-Murillo, J.M., Rodrguez, A.A., Baca-Garca, E.: Visualization and prediction of disease interactions with continuous-time hidden Markov models. In: Advances in Neural Information Processing Systems (2011)
Liu, Y., Ishikawa, H., Chen, M., et al.: Longitudinal modeling of glaucoma progression using 2-dimensional continuous-time hidden Markov model. In: Proc. Medical Image Computing and Computer-Assisted Intervention (MICCAI), pp. 444–451 (2013)
Liu, Y.Y., Li, S., Li, F., Song, L., Rehg, J.M.: Efficient learning of continuous-time hidden Markov models for disease progression. In: Proc. Twenty-Ninth Annual Conference on Neural Information Processing Systems (NIPS 15). Montreal, Canada (2015)
McGibbon, R.T., Pande, V.S.: Efficient maximum likelihood parameterization of continuous-time Markov processes. The Journal of Chemical Physics 143(3), 034,109 (2015)
Metzner, P., Horenko, I., Schütte, C.: Generator estimation of Markov jump processes. Journal of Computational Physics 227, 353–375 (2007)
Metzner, P., Horenko, I., Schütte, C.: Generator estimation of Markov jump processes based on incomplete observations nonequidistant in time. Physical Review E 76(066702) (2007)
Moler, C., Van Loan, C.: Nineteen dubious ways to compute the exponential of a matrix. SIAM Review 20(4), 801–836 (1978)
Moler, C., Van Loan, C.: Nineteen dubious ways to compute the exponential of a matrix, twenty-five years later. SIAM Review 45(1), 3–49 (2003)
Nodelman, U., Shelton, C.R., Koller, D.: Expectation maximization and complex duration distributions for continuous time Bayesian networks. In: Proc. Uncertainty in AI (UAI 05) (2005)
Osborne, E.: On pre-conditioning of matrices. Journal of the ACM (JACM) 7(4), 338–345 (1960)
Rabinar, L.R.: A tutorial on Hidden Markov Models and selected applications in speech recognition. Proceedings of the IEEE 77(2) (1989)
Ross, S.M.: Stochastic Processes. John Wiley, New York (1983)
Tataru, P., Hobolth, A.: Comparison of methods for calculating conditional expectations of sufficient statistics for continuous time Markov chains. BMC Bioinformatics 12(465) (2011)
The Alzheimer’s Disease Neuroimaging Initiative: http://adni.loni.usc.edu
Van Loan, C.: Computing integrals involving the matrix exponential. IEEE Trans. Automatic Control 23, 395–404 (1978)
Wang, X., Sontag, D., Wang, F.: Unsupervised learning of disease progression models. Proceeding KDD 4(1), 85–94 (2014)
Wollstein, G., Kagemann, L., Bilonick, R., et al.: Retinal nerve fibre layer and visual function loss in glaucoma: the tipping point. Br J Ophthalmol 96(1), 47–52 (2012)
Acknowledgements
Portions of this work were supported in part by NIH R01 EY13178-15 and by grant U54EB020404 awarded by the National Institute of Biomedical Imaging and Bioengineering through funds provided by the Big Data to Knowledge (BD2K) initiative (www.bd2k.nih.gov). The research was also supported in part by NSF/NIH BIGDATA 1R01GM108341, ONR N00014-15-1-2340, NSF IIS-1218749, NSF CAREER IIS-1350983, and funding from the Georgia Tech Executive Vice President of Research Office and the Center for Computational Health.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Appendix: Derivation of Vectorized Eigen
Appendix: Derivation of Vectorized Eigen
In [20, 21], it is stated without proof that the naïve Eigen is equivalent to Vectorized Eigen. Here we present the derivation. Let
where the symmetric matrix Ψ(t) = [Ψ pq (t)] p, q ∈ S is defined as:
Letting V = U −1, this is equivalent to
To see why, first, note that for the outer product,
Then
So that
as desired.
Rights and permissions
Copyright information
© 2017 Springer International Publishing AG
About this chapter
Cite this chapter
Liu, YY., Moreno, A., Li, S., Li, F., Song, L., Rehg, J.M. (2017). Learning Continuous-Time Hidden Markov Models for Event Data. In: Rehg, J., Murphy, S., Kumar, S. (eds) Mobile Health. Springer, Cham. https://doi.org/10.1007/978-3-319-51394-2_19
Download citation
DOI: https://doi.org/10.1007/978-3-319-51394-2_19
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-51393-5
Online ISBN: 978-3-319-51394-2
eBook Packages: Computer ScienceComputer Science (R0)