impute: Multiple imputation for data analysis via mixtures

In burrisk/midamix: Bayesian multiple imputation via mixtures

Description

Uses a Dirichlet Process Mixture Transformation Model to sample missing data values from the corresponding posterior distribution.

Usage

impute(data, imputations = 10, max_clusters = 15, n_iter = 1000,
  burnin = 100, validator = NULL, cap = NULL, seed = 1)

Arguments

data	A data frame, consisting of numeric, integer, or ordered factor columns.
imputations	The number of imputed datasets to return, defaults to 10.
max_clusters	The maximum number of clusters for the mixture model.
n_iter	Number of iterations for the MCMC sampler.
burnin	Number of iterations for initial burn-in period.
validator	A function that takes in an observation and determines whether it is feasible.
cap	Maximum number of rejected proposals allowed for the constrained sampler
seed	Random seed.

Value

A tibble consisting of multiply imputed data sets.

burrisk/midamix documentation built on June 1, 2019, 12:49 p.m.