Joshua Warren, PhD

This package implements a hierarchical Bayesian meta analysis/regression for areal spatial data and accounts for spatial correlation in the data in multiple ways. There is also the option for a more traditional non-spatial analysis. The method relies on the conditional autoregressive model and is fit using Markov chain Monte Carlo sampling techniques. Please see the "SpMeta_Model_Details" and "SpMeta_Example" folders for more specific information regarding the statistical model and package use details, respectively.

Faculty: Joshua Warren, PhD

Download: GitHub / SpMeta package

Platform: R

Code for implementing RTS_P and RTS_NB from: Warren JL, Prunas O, Paltiel AD, Thornhill T, and Gonsalves GS (2024). Integrating testing volume into bandit algorithms for infectious disease surveillance. Journal of the Royal Statistical Society: Series A, In Press.

Faculty: Joshua Warren, PhD

Download: GitHub / Bandits package

Platform: R

Reference: doi.org (Bandits)

This package allows for univariate and multivariate kernel density estimation prior distributions, defined by posterior predictive samples collected in a first stage exposure model, to be assigned to exposures within a health outcome regression analysis (continuous, binary, and count models included), and is fit using Markov chain Monte Carlo sampling techniques. Please see the "KDExp_Model_Details" and "KDExp_Example" folders for more specific information regarding the statistical model and package use details, respectively.

Faculty: Joshua Warren, PhD

Download: GitHub / KDExp package

Platform: R

Reference: doi.org (KDExp)

This package implements a hierarchical Bayesian logistic regression analysis to estimate a spatial change point corresponding to spillover from a point source while simultaneously accounting for small scale spatial variability via spatially correlated random effects (spherical covariance function). A similar method was used to estimated spillover of multidrug-resistant tuberculosis from a prison in Lima Peru (Warren et al. 2018). Please see the "Spillover_Model_Details" and "Spillover_Example" folders for more specific information regarding the statistical model and package use details, respectively.

Faculty: Joshua Warren, PhD; Ted Cohen, DPH, MD, MPH

Download: GitHub / Spillover package

Platform: R

Reference: doi.org (Spillover)

This package implements a hierarchical Bayesian logistic regression analysis to estimate spatially varying critical windows of susceptibility corresponding to exposure from a single time-varying covariate. The method relies on a spatiotemporally structured Gaussian process and is fit using Markov chain Monte Carlo sampling techniques. Please see the "SpGPCW_Example" folder for more specific information regarding package use details.

Faculty: Joshua Warren, PhD

Download: GitHub / SpGPCW package

Platform: R

Reference: doi.org (SpGPCW)

This package implements a hierarchical Bayesian logistic regression analysis to estimate spatially varying critical windows of susceptibility corresponding to exposure from a single time-varying covariate. The method relies on a spatiotemporally structured Gaussian process and is fit using Markov chain Monte Carlo sampling techniques. Please see the "SpGPCW_Example" folder for more specific information regarding package use details.

Faculty: Joshua Warren, PhD

Download: GitHub / SpGPCW package

Platform: R

Reference: doi.org (SpGPCW)

This package implements a Directionally-Varying Change Points (DVCP) model that aims to estimate the magnitude of the impact of a point source as well as its range of influence across the spatial domain. DVCP includes a Gaussian process with directionally-defined correlation structure nested within a change point framework to introduce unique change point parameters in every direction extending from the point source. The Gaussian predictive process approximation is used to facilitate model fitting for large datasets, and DVCP is fit using Markov chain Monte Carlo sampling techniques. Please see the "DVCP_Model_Details" and "DVCP_Example" folders for more specific information regarding the statistical model and package use details, respectively.

Faculty: Joshua Warren, PhD

Download: GitHub / DVCP package

Platform: R

Reference: doi.org (DVCP)

This package implements a hierarchical Bayesian logistic regression analysis to estimate critical windows of susceptibility corresponding to exposure from a single time-varying covariate. The method relies on a temporally structured Gaussian process and is fit using Markov chain Monte Carlo sampling techniques. Similar methods have been used in Warren et al. (2012) to identify vulnerable periods of pregnancy with respect to ambient air pollution exposure. Please see the "GPCW_Model_Details" and "GPCW_Example" folders for more specific information regarding the statistical model and package use details, respectively.

Faculty: Joshua Warren, PhD

Download: GitHub / GPCW package

Platform: R

Reference: doi.org (GPCW)

This package implements a hierarchical Bayesian logistic regression analysis to identify/estimate critical windows of susceptibility corresponding to exposure from a single time-varying covariate. The method involves temporally smoothed Bayesian variable selection, with correlated Gaussian process smoothness in the risk and variable selection parameters, and is fit using Markov chain Monte Carlo sampling techniques. Please see the "CWVS_Model_Details" and "CWVS_Example" folders for more specific information regarding the statistical model and package use details, respectively.

Faculty: Joshua Warren, PhD

Download: GitHub / CWVS package

Platform: R

Reference: doi.org (CWVS)

This package implements hierarchical Bayesian regression models to analyze factors associated with genetic similarity between pairs of individuals while accounting for multiple sources of correlation, and is fit using Markov chain Monte Carlo sampling techniques. Please see the "GenePair_Model_Details" and "GenePair_Example" folders for more specific information regarding the statistical model and package use details, respectively.

Faculty: Joshua Warren, PhD; Ted Cohen, DPH, MD, MPH

Download: GitHub / GenePair package

Platform: R

Reference: doi.org (GenePair)

This package implements a hierarchical Bayesian Poisson regression analysis with nonparametric spatially correlated random effects useful for detecting boundaries in spatial data. The method relies on a discrete areal kernel function introduced within the kernel stick-breaking framework and is fit using Markov chain Monte Carlo sampling techniques. Please see the "KSBound_Model_Details" and "KSBound_Example" folders for more specific information regarding the statistical model and package use details, respectively.

Faculty: Joshua Warren, PhD; Nicole Deziel, PhD, MHS

Download: GitHub / KSBound package

Platform: R

Reference: doi.org (KSBound)

This package implements a hierarchical Bayesian spatially-varying (and spatiotemporally-varying) distributed lag regression analysis to predict ambient air pollution concentrations at new spatial locations and times. The method uses lagged average gridded air pollution estimates as predictor information and is fit using Markov chain Monte Carlo sampling techniques. Please see the "DLfuse_Model_Details" and "DLfuse_Examples" folders for more specific information regarding the statistical model and package use details, respectively.

Faculty: Joshua Warren, PhD; Michelle L. Bell, PhD

Download: GitHub / DLfuse package

Platform: R

Reference: doi.org (DLfuse)