Bayesian Bootstrap Predictive Mean Matching - Multiple and
Single Imputation for Discrete Data
Included are two variants of Bayesian Bootstrap
Predictive Mean Matching to multiply impute missing data. The
first variant is a variable-by-variable imputation combining
sequential regression and Predictive Mean Matching (PMM) that
has been extended for unordered categorical data. The Bayesian
Bootstrap allows for generating approximately proper multiple
imputations. The second variant is also based on PMM, but the
focus is on imputing several variables at the same time. The
suggestion is to use this variant, if the missing-data pattern
resembles a data fusion situation, or any other
missing-by-design pattern, where several variables have
identical missing-data patterns. Both variants can be run as
'single imputation' versions, in case the analysis objective is
of a purely descriptive nature.
Changes in version BaBooN 0.2-0 (2015-06-15)
- BBPMM.row: Implementation of Predictive Mean Matching via RcppArmadillo. stepmod replaces the old argument stepwise. Changes in code structure. Additional return values.
- BBPMM: maxit.multi and maxit.glm replace the old argument maxit. stepmod replaces the old argument stepwise. Changes in code structure. Additional return values.
- rowimpprep: Additional return values.
- DESCRIPTION: Forces byte compilation.
- NAMESPACE: exports more S3 methods, imports only for specific functions
- dmi (new): Auxiliary function for BBPMM. Creates Data monotonicity index for missing values in order to regulate the numbers of iterations (if BBPMM argument nIter is set to "autolin").
- impdiagnosticconversion (new): Conversion from BaBooN's BBPMM to coda's mcmc or mcmc.list object, or to mice's mids object.