R Programming/Multiple Imputation

Multiple imputation of missing data generally includes two steps. First, an imputation step which results in multiple complete datasets. Second, combining the results obtained by applying the chosen technique on each separate dataset. The packages needed for these two steps are not necessary the same, but can be.


 * "mi" package Andrew Gelman Blog Entry on the "mi" package
 * "mice" package.
 * Amelia
 * mitools by Thomas Lumley
 * survey and Zelig have support for multiple imputated datasets.