Latent
Markov and related models
for longitudinal categorical
data
Aim of this page is to help the diffusion of models for longitudinal
categorical data that are formulated through a Markov chain latent process
or a similar latent process that follows, for instance, an AR(1). These
models have a structure similar to that of hidden Markov models for
time-series data. From the methodological point of view, they are of
interest for researchers in Statistics and Econometrics. Moreover, they
may be fruitfully applied in the analysis of longitudinal datasets in
Economics, Medicine, and Social Sciences, even from a perspective causal
inference.
With respect to standard model for the analysis of longitudinal data, latent
Markov and related models allow us:
- to account for unobserved heterogeneity (additional to the
heterogeneity that may be explained by the observable covariates) in a
dynamic fashion;
- to summarize several outcomes observed at the same occasion into
a single variable (e.g., quality of life), the evolution of which on
time may be studied even depending on individual covariates.
The web page intends to help the diffusion of latent
Markov and related models by:
- listing researchers actively working on these models;
- listing relevant articles, books, technical reports, and software
in the field;
- reporting and supporting conferences and events that may be of
interest for presenting papers in the field.
The webpage is maintained by Francesco
Bartolucci, Department of Statistics, University of Perugia (IT).
Please contact him by email (bart@stat.unipg)
to be added among researchers working on the field, to list relevant
papers, or include advertise to international conferences and events.
- Main references
- Researchers
- Publications
- Techinical Reports
- Presentations
Main references
Wiggins, L.M (1973). Panel Analysis:
Latent probability models for attitude and behaviour processes.
Elsevier, Amsterdam.
Bartolucci, F., Farcomeni, A. and Pennoni, F. (2012), Latent
Markov Models for Longitudinal Data, Chapman and Hall/CRC press.
Up-to-date bibliografy
Researchers
Marco
Alfò, Department of Statistical Sciences, University of Rome - La
Sapienza (IT)
Silvia
Bacci, Department of Economics, Finance and Statistics, University
of Perugia (IT)
Francesco
Bartolucci, Department of Economics, Finance and
Statistics, University of Perugia (IT)
Silvia Pandolfi, Department of Economics, Finance and
Statistics, University of Perugia (IT)
Publications
2011
Bartolucci, F., Pennoni, F. and Vittadini, G. (2011), Assessment
of school performance through a multilevel latent Markov Rasch model,
Journal of Educational and Behavioral
Statistics, 36, pp. 491-522.
2010
Alfò, M. and Maruotti, A. (2010), Two-part regression models for
longitudinal zero-inflated count data, Canadian
Journal of Statistics, 38,
pp. 197–216.
2009
Alfò, M. and Maruotti, A. (2009), A hierarchical
Model for Time Dependent Multivariate Longitudinal Data. In F. Palumbo,
N.C. Lauro, M. Greenacre (eds.), Data
Analysis and Classification: from the exploratory to the confirmatory
approach, Springer Verlag, Heidelberg, pp. 271–279.
Techinical Reports
2012
Bacci, S., Pandolfi, S. and Pennoni, F. (2012), A
comparison of some criteria for states selection in the latent Markov model
for longitudinal data, arXiv:1212.0352
2011
Bartolucci, F., Bacci, S., and Pennoni, F. (2011), Mixture latent
autoregressive models for longitudinal data, arXiv:1108.1498v1
2010
Bartolucci, F. and Pandolfi, S. (2010). Bayesian inference for a class
of latent Markov models for categorical longitudinal data. arXiv:
1101.0391v2
Presentations
2012
Bacci, S., Bartolucci, F., Pandolfi, S. and Pennoni, F. (2012), A
comparisons of some criteria for states selection of the latent Markov
model for longitudinal data, MBC2 -
Workshop on Model Based Clustering and Classification - Catania
(IT) - September 6-7, 2012.
Montanari, G. E. and Pandolfi, S. (2012), Evaluation of Nursing Homes
Using an Extended Latent Markov Model, JCS
– CLADAG 2012 - Anacapri - Settembre 2010.
2010
Farcomeni, A. and Pandolfi, S. (2010) Bayesian multivariate latent Markov
models with an unknown number of regimes. XLV
conference Scientifica della Società Italiana di Statistica (SIS),
Padova, 16-18 giugno 2010.