Mixed Hidden Markov Model for Heterogeneous Longitudinal Data with Missingness and Errors in the Outcome Variable

(2014) - Dominique Dedieu, Cyrille Delpierre, Sébastien Gadat, Thierry Lang, Benoît Lepage, Nicolas Savy

Journal:
Link::
DOI::
Links::
Tags:: #paper #NCDS #MarkovModel #MissingData
Cite Key:: [@dedieuMixedHiddenMarkov2014]

Abstract

Analysing longitudinal declarative data raises many difficulties, such as the processing of errors and missingness in the outcome variable. Moreover, long-term monitored cohorts (commonly encountered in life-course epidemiology) may reveal a problem of time heterogeneity, especially regarding the way subjects respond to the investigator. We propose a Mixed Hidden Markov Model which considers several causes of randomness in response and also enables the effect of a past health outcome to act on present responses through a memory state. Hence, we take into account both errors and missing responses, time heterogeneity, and retrospective questions. We thus propose a Stochastic Expectation Maximization algorithm (SEM), which is less time-consuming than usual EM algorithms to perform the estimation of the parameters of our MHMM.

Notes

“We propose a Mixed Hidden Markov Model which considers several causes of randomness in response and also enables the effect of a past health outcome to act on present responses through a memory state. Hence, we take into account both errors and missing responses, time heterogeneity, and retrospective questions. We thus propose a Stochastic Expectation Maximization algorithm (SEM), which is less time-consuming than usual EM algorithms to perform the estimation of the parameters of our MHMM” (Dedieu et al., 2014, p. 73)