Browsing by Author "Shin, Hwasoo"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Bayesian Factor Models for Clustering and Spatiotemporal Analysis
    Shin, Hwasoo (Virginia Tech, 2024-05-28)
    Multivariate data is prevalent in modern applications, yet it often presents significant analytical challenges. Factor models can offer an effective tool to address issues associated with large-scale datasets. In this dissertation, we propose two novel Bayesian factors models. These models are designed to effectively reduce the dimensionality of the data, as the number of latent factors is typically much smaller than that of the observation vectors. Therefore, our proposed models can achieve substantial dimension reduction. Our first model is for spatiotemporal areal data. In this case, the region of interest is divided into subregions, and at each time point, there is one univariate observation per subregion. Our model writes the vector of observations at each time point in a factor model form as the product of a vector of factor loadings and a vector of common factors plus a vector of error. Our model assumes that the common factor evolves through time according to a dynamic linear model. To represent the spatial relationships among subregions, each column of the factor loadings matrix is assigned intrinsic conditional autoregressive (ICAR) priors. Therefore, we call our approach the Dynamic ICAR Spatiotemporal Factor Models (DIFM). Our second model, Bayesian Clustering Factor Model (BCFM) assumes latent factors and clusters are present in the data. We apply Gaussian mixture models on common factors to discover clusters. For both models, we develop MCMC to explore the posterior distribution of the parameters. To select the number of factors and, in the case of clustering methods, the number of clusters, we develop model selection criteria that utilize the Laplace-Metropolis estimator of the predictive density and BIC with integrated likelihood.
  • Thumbnail Image
    Long-term recovery from opioid use disorder: recovery subgroups, transition states and their association with substance use, treatment and quality of life
    Craft, William H.; Shin, Hwasoo; Tegge, Allison N.; Keith, Diana R.; Athamneh, Liqa N.; Stein, Jeffrey S.; Ferreira, Marco A. R.; Chilcoat, Howard D.; Le Moigne, Anne; DeVeaugh-Geiss, Angela; Bickel, Warren K. (Wiley, 2022-12)
    Background and AimsLimited information exists regarding individual subgroups of recovery from opioid use disorder (OUD) following treatment and how these subgroups may relate to recovery trajectories. We used multi-dimensional criteria to identify OUD recovery subgroups and longitudinal transitions across subgroups. Design, Setting and ParticipantsIn a national longitudinal observational study in the United States, individuals who previously participated in a clinical trial for subcutaneous buprenorphine injections for treatment of OUD were enrolled and followed for an average of 4.2 years after participation in the clinical trial. MeasurementsWe identified recovery subgroups based on psychosocial outcomes including depression, opioid withdrawal and pain. We compared opioid use, treatment utilization and quality of life among these subgroups. FindingsThree dimensions of the recovery process were identified: depression, opioid withdrawal and pain. Using these three dimensions, participants were classified into four recovery subgroups: high-functioning (minimal depression, mild withdrawal and no/mild pain), pain/physical health (minimal depression, mild withdrawal and moderate pain), depression (moderate depression, mild withdrawal and mild/moderate pain) and low-functioning (moderate/severe withdrawal, moderate depression and moderate/severe pain). Significant differences among subgroups were observed for DSM-5 criteria (P < 0.001) and remission status (P < 0.001), as well as with opioid use (P < 0.001), treatment utilization (P < 0.001) and quality of life domains (physical health, psychological, environment and social relationships; Ps < 0.001, Cohen's fs >= 0.62). Recovery subgroup assignments were dynamic, with individuals transitioning across subgroups during the observational period. Moreover, the initial recovery subgroup assignment was minimally predictive of long-term outcomes. ConclusionsThere appear to be four distinct subgroups among individuals in recovery from OUD. Recovery subgroup assignments are dynamic and predictive of contemporaneous, but not long-term, substance use, substance use treatment utilization or quality of life outcomes.