imp_cat_multi: The function for hierarchical imputation of categorical...

In matthiasspeidel/hmi: Hierarchical Multiple Imputation

Description

The function is called by the wrapper and relies on MCMCglmm.
While in the single level function (imp_cat_single) we used regression trees to impute data, here we run a multilevel multinomial model. The basic idea is that for each category of the target variable (expect the reference category) a own formula is set up, saying for example that the chances to end up in category j increase with increasing X5. So there is a own regression coefficient beta_5_j present. In a multilevel setting, this regression coefficient beta_5_j might be different for different clusters so for cluster 1 it would be beta_5_j_1 = beta_5_j + u_5_1 and for cluster 27 beta_5_j_27 = beta_5_j + u_5_27. This also leads to own random effect covariance matrices for each category. Or, if you want to have all random effect variance parameters in one matrix: a (very large) matrix where not for example only the random intercepts variance and random slopes variance and their covariance is present. Instead there is even a covariance between the random slopes in category 2 and the random intercepts in category 4. For simplicity these covariances are set to be 0.

Usage

imp_cat_multi(y_imp_multi, X_imp_multi, Z_imp_multi, clID, model_formula,
  M = 10, nitt = 3000, thin = 10, burnin = 1000)

Arguments

y_imp_multi	A Vector with the variable to impute.
X_imp_multi	A data.frame with the fixed effects variables.
Z_imp_multi	A data.frame with the random effects variables.
clID	A vector with the cluster ID.
model_formula	A formula used for the analysis model.
M	An integer defining the number of imputations that should be made.
nitt	An integer defining number of MCMC iterations (see MCMCglmm).
thin	An integer defining the thinning interval (see MCMCglmm).
burnin	An integer defining the percentage of draws from the gibbs sampler that should be discarded as burn in (see MCMCglmm).

Value

A n x M matrix. Each column is one of M imputed y-variables.

matthiasspeidel/hmi documentation built on Aug. 18, 2020, 4:37 p.m.