A semiparametric approach to hidden Markov models under longitudinal observations
A semiparametric approach to hidden Markov models under longitudinal observations
We propose a hidden Markov model for longitudinal count data where sources of unobserved heterogeneity arise, making data overdispersed. The observed process, conditionally on the hidden states, is assumed to follow an inhomogeneous Poisson kernel, where the unobserved heterogeneity is modeled in a generalized linear model (GLM) framework by adding individual-specific random effects in the link function.
Due to the complexity of the likelihood within the GLM framework, model parameters may be estimated by numerical maximization of the log-likelihood function or by simulation methods; we propose a more flexible approach based on the Expectation Maximization (EM) algorithm. Parameter estimation is carried out using a non-parametric maximum likelihood (NPML) approach in a finite mixture context. Simulation results and two empirical examples are provided.
hidden markov models, longitudinal data, mixed hidden markov models, random effects, npml
381-393
Maruotti, Antonello
7096256c-fa1b-4cc1-9ca4-1a60cc3ee12e
Ryden, Tobias
a4750d5f-81b5-4a62-b74b-7a191f84153d
2009
Maruotti, Antonello
7096256c-fa1b-4cc1-9ca4-1a60cc3ee12e
Ryden, Tobias
a4750d5f-81b5-4a62-b74b-7a191f84153d
Maruotti, Antonello and Ryden, Tobias
(2009)
A semiparametric approach to hidden Markov models under longitudinal observations.
Statistics and Computing, 19 (4), .
(doi:10.1007/s11222-008-9099-2).
Abstract
We propose a hidden Markov model for longitudinal count data where sources of unobserved heterogeneity arise, making data overdispersed. The observed process, conditionally on the hidden states, is assumed to follow an inhomogeneous Poisson kernel, where the unobserved heterogeneity is modeled in a generalized linear model (GLM) framework by adding individual-specific random effects in the link function.
Due to the complexity of the likelihood within the GLM framework, model parameters may be estimated by numerical maximization of the log-likelihood function or by simulation methods; we propose a more flexible approach based on the Expectation Maximization (EM) algorithm. Parameter estimation is carried out using a non-parametric maximum likelihood (NPML) approach in a finite mixture context. Simulation results and two empirical examples are provided.
This record has no associated files available for download.
More information
e-pub ahead of print date: 27 September 2008
Published date: 2009
Keywords:
hidden markov models, longitudinal data, mixed hidden markov models, random effects, npml
Organisations:
Statistics, Statistical Sciences Research Institute
Identifiers
Local EPrints ID: 345968
URI: http://eprints.soton.ac.uk/id/eprint/345968
ISSN: 0960-3174
PURE UUID: 55bf4022-29be-4fba-a28b-8bec9eefa9ef
Catalogue record
Date deposited: 10 Dec 2012 14:00
Last modified: 14 Mar 2024 12:31
Export record
Altmetrics
Contributors
Author:
Antonello Maruotti
Author:
Tobias Ryden
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics