Advanced

A semiparametric approach to hidden Markov models under longitudinal observations

Maruotti, Antonello and Rydén, Tobias LU (2009) In Statistics and Computing 19(4). p.381-393
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... (More)
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. (Less)
Please use this url to cite or link to this publication:
author
organization
publishing date
type
Contribution to journal
publication status
published
subject
keywords
NPML, Random effects, Mixed hidden Markov models, Hidden Markov models, Longitudinal data
in
Statistics and Computing
volume
19
issue
4
pages
381 - 393
publisher
Springer
external identifiers
  • wos:000271738000003
  • scopus:70450284761
ISSN
0960-3174
DOI
10.1007/s11222-008-9099-2
language
English
LU publication?
yes
id
f0d02331-d838-40d7-bb99-65a033c980e7 (old id 1518665)
date added to LUP
2009-12-28 14:42:20
date last changed
2017-12-10 04:20:17
@article{f0d02331-d838-40d7-bb99-65a033c980e7,
  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.},
  author       = {Maruotti, Antonello and Rydén, Tobias},
  issn         = {0960-3174},
  keyword      = {NPML,Random effects,Mixed hidden Markov models,Hidden Markov models,Longitudinal data},
  language     = {eng},
  number       = {4},
  pages        = {381--393},
  publisher    = {Springer},
  series       = {Statistics and Computing},
  title        = {A semiparametric approach to hidden Markov models under longitudinal observations},
  url          = {http://dx.doi.org/10.1007/s11222-008-9099-2},
  volume       = {19},
  year         = {2009},
}