Multiple Imputation Using Weighted Quantile Sum Regression
Performing Individual Chemical Analysis
Finding WQS Coefficients
Combining AICs
Performing Many WQS Regressions
Formula for WQS Regression
Weighted Quantile Sum (WQS) Regression
Bootstrapping Imputation for Many Chemicals
Lubin et al. 2004: Bootstrapping Imputation for One Chemical
Multivariate Bayesian Imputation
Imputing by Substitution
Univariate Bayesian Imputation
Making Quantiles of Correlated Index
Histograms of the Weights, Beta1, and WQS using ggplot
Pooling Multiple Imputation Results
Prints the fitted WQS model along with the mean weights.
The miWQS package handles the uncertainty due to below the detection limit in a correlated component mixture problem. Researchers want to determine if a set/mixture of continuous and correlated components/chemicals is associated with an outcome and if so, which components are important in that mixture. These components share a common outcome but are interval-censored between zero and low thresholds, or detection limits, that may be different across the components. This package applies the multiple imputation (MI) procedure to the weighted quantile sum regression (WQS) methodology for continuous, binary, or count outcomes (Hargarten & Wheeler (2020) <doi:10.1016/j.envres.2020.109466>). The imputation models are: bootstrapping imputation (Lubin et.al (2004) <doi:10.1289/ehp.7199>), univariate Bayesian imputation (Hargarten & Wheeler (2020) <doi:10.1016/j.envres.2020.109466>), and multivariate Bayesian regression imputation.