As stated in the paper missForest already contains a "quasi-multiple imputation" scheme. By generating many trees in the random forest, we do get a population of imputed values and can extract a standard deviation on these (in the continuous case). This would allow for correcting the deflated sd in the imputed data as described in Schafer, 1997.
This would ultimately allow to subsequently use the imputed data for statistical inference.
As stated in the paper missForest already contains a "quasi-multiple imputation" scheme. By generating many trees in the random forest, we do get a population of imputed values and can extract a standard deviation on these (in the continuous case). This would allow for correcting the deflated sd in the imputed data as described in Schafer, 1997.
This would ultimately allow to subsequently use the imputed data for statistical inference.