carpenitoThomas/misl: Multiple Imputation by Super Learning

The MISL algorithm is similar to other full conditional specification methods in that a series of m unique datasets are generated through convergence of multiple iterations however the true predictive power of MISL is highlighted with the advantage of not needing to make explicit modeling assumptions about the distribution of the data. This allows for variability in imputations to come from the variation in cross validation, not from explicitly modeling posterior distributions.

install.packages("remotes")
remotes::install_github("carpenitoThomas/misl")