Bayesian analysis for mixture of latent variable hidden Markov models with multivariate longitudinal data

Xia, Ye-Mao; Tang, Niansheng

doi:10.1016/j.csda.2018.08.004

Cited by 16 publications

(8 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…both continuous and discrete). For the former, example include multivariate model-based clustering using mixed effect models for continuous markers (Marshall et al, 2006;Villarroel et al, 2009;Proust-Lima et al, 2017), multivariate functional clustering for continuous (Jacques and Preda, 2014b) or discrete markers (Lim et al, 2020), Bayesian finite mixtures for continuous markers (Leiby et al, 2009;Frühwirth-Schnatter and Pyne, 2010) and hidden Markov models for continuous markers (Xia and Tang, 2019;Lu and Lou, 2021c). For the latter, examples include group-based multi-trajectory analysis (Nagin et al, 2018), which is an extension of the group-based single trajectory analysis (Nagin, 1999), and Bayesian mixture modeling for discrete and continuous longitudinal data (BMM) (Komárek et al, 2013(Komárek et al, , 2014.…”

Section: Review Of Existing Statistical Methodsmentioning

confidence: 99%

A Joint Modeling Approach for Clustering Mixed-Type Multivariate Longitudinal Data: Application to the CHILD Cohort Study

Tan¹,

Shen²,

Subbarao³

et al. 2022

Preprint

View full text Add to dashboard Cite

In epidemiological and clinical studies, identifying patients' phenotypes based on longitudinal profiles is critical to understanding the disease's developmental patterns. The current study was motivated by data from a Canadian birth cohort study, the CHILD Cohort Study. Our goal was to use multiple longitudinal respiratory traits to cluster the participants into subgroups with similar longitudinal respiratory profiles in order to identify clinically relevant disease phenotypes. To appropriately account for distinct structures and types of these longitudinal markers, we proposed a novel joint model for clustering mixed-type (continuous, discrete and categorical) multivariate longitudinal data. We also developed a Markov Chain Monte Carlo algorithm to estimate the posterior distribution of model parameters. Analysis of the CHILD Cohort data and simulated data were presented and discussed. Our study demonstrated that the proposed model serves as a useful analytical tool for clustering multivariate mixed-type longitudinal data. We developed an R package BCClong to implement the proposed model efficiently.

show abstract

Section: Review Of Existing Statistical Methodsmentioning

confidence: 99%

A Joint Modeling Approach for Clustering Mixed-Type Multivariate Longitudinal Data: Application to the CHILD Cohort Study

Tan¹,

Shen²,

Subbarao³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Note that for the moment, we assume that the elements in boldθ are all free. If some constraints on the parameters are imposed, the method of (Xia and Tang (2019)) can be used to deal with this issue. More general treatment on this issue can refer to, for example, (Lee (2007)).…”

Section: Bayesian Inferencementioning

confidence: 99%

Bayesian analysis of two-part nonlinear latent variable model: Semiparametric method

Gou

Xia

Jiang

2021

Statistical Modelling

Self Cite

View full text Add to dashboard Cite

Two-part model (TPM) is a widely appreciated statistical method for analyzing semi-continuous data. Semi-continuous data can be viewed as arising from two distinct stochastic processes: one governs the occurrence or binary part of data and the other determines the intensity or continuous part. In the regression setting with the semi-continuous outcome as functions of covariates, the binary part is commonly modelled via logistic regression and the continuous component via a log-normal model. The conventional TPM, still imposes assumptions such as log-normal distribution of the continuous part, with no unobserved heterogeneity among the response, and no collinearity among covariates, which are quite often unrealistic in practical applications. In this article, we develop a two-part nonlinear latent variable model (TPNLVM) with mixed multiple semi-continuous and continuous variables. The semi-continuous variables are treated as indicators of the latent factor analysis along with other manifest variables. This reduces the dimensionality of the regression model and alleviates the potential multicollinearity problems. Our TPNLVM can accommodate the nonlinear relationships among latent variables extracted from the factor analysis. To downweight the influence of distribution deviations and extreme observations, we develop a Bayesian semiparametric analysis procedure. The conventional parametric assumptions on the related distributions are relaxed and the Dirichlet process (DP) prior is used to improve model fitting. By taking advantage of the discreteness of DP, our method is effective in capturing the heterogeneity underlying population. Within the Bayesian paradigm, posterior inferences including parameters estimates and model assessment are carried out through Markov Chains Monte Carlo (MCMC) sampling method. To facilitate posterior sampling, we adapt the Polya-Gamma stochastic representation for the logistic model. Using simulation studies, we examine properties and merits of our proposed methods and illustrate our approach by evaluating the effect of treatment on cocaine use and examining whether the treatment effect is moderated by psychiatric problems.

show abstract

“…Ref. [17] incorporated a two-part regression model into a general latent variable model framework and analyzed the internal relationships between multiple factors longitudinally. These methods have brought significant attention to two-part models in behavioral science, economics, psychology, and medicine in recent years: see, for example [14,18,19], and references therein for further developments of two-part models.…”

Section: Introductionmentioning

confidence: 99%

Bayesian Feature Extraction for Two-Part Latent Variable Model with Polytomous Manifestations

Zhang,

Xia

2024

Preprint

View full text Add to dashboard Cite

Semi-continuous data are very common in the social science and economics. In this paper, a Bayesian variable selection procedure is developed to assess the influence of exogenous factors including observed and unobserved on the semi-continuous data. Our formulation is based on the two-part latent variable model with polytomous response. We consider two schemes for the penalties of regression coefficients and factor loadings: the Bayesian spike and slab bimodal prior and the Bayesian lasso prior. Within the Bayesian framework, we implement Markov Chains Monte Carlo sampling method to conduct posterior inference. To facilitate posterior sampling, we recast the logistic model in part one as the norm-like mixture model. Gibbs sampler is designed to draw observations from the posterior. Our empirical results show that with suitable hyper-parameters, the spike and slab bimodal method slightly outperforms the Bayesian lasso in the current analysis. Finally, a real example related to the China household finance survey is analyzed to illustrate the application of the methodology.. .

show abstract

Bayesian analysis for mixture of latent variable hidden Markov models with multivariate longitudinal data

Cited by 16 publications

References 45 publications

A Joint Modeling Approach for Clustering Mixed-Type Multivariate Longitudinal Data: Application to the CHILD Cohort Study

A Joint Modeling Approach for Clustering Mixed-Type Multivariate Longitudinal Data: Application to the CHILD Cohort Study

Bayesian analysis of two-part nonlinear latent variable model: Semiparametric method

Bayesian Feature Extraction for Two-Part Latent Variable Model with Polytomous Manifestations

Contact Info

Product

Resources

About