-1

I'm going to compare my model with the EMB method in Amelia package. I read the article, but it does not mentioned specifically to call the EMB method from Amelia.

I got two questions:

  1. how to call EMB from Amelia?
  2. Does it correct if i want to impute the missing data in Amelia,

The code is this:

Completed_data<-amelia(XNanData,m=3,p2s=0)
dataI=Completed_data$imputations[[3]]

So,is the dataI considered as the imputed data in Amelia II? From my understanding, Multiple imputation consist of three steps:Imputation, Analysis and Pool.

Is dataI from above considered the pooled result or i need to call another function in Amelia for the pooled result? I'm still new in MI.

StupidWolf
  • 45,075
  • 17
  • 40
  • 72
amjay
  • 11
  • 6
  • Welcome to SO. Please provide a [reproducible example](http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example), especially a reproducible dataset. It would also be helpful to provide a link to the "EMB" algorithm for those who are unfamiliar with it. – alexwhitworth Jan 23 '18 at 19:04

1 Answers1

1

The default algorithm used in Amelia II (library(Amelia)) is the EMB algorithm (EM with bootstrapping). Any call to Amelia::amelia will use EMB.

As mentioned in the manual you link to:

The EM algorithm (Dempster, Laird and Rubin, 1977) is a simple computational approach to finding the mode of the posterior. Our EMB algorithm combines the classic EM algorithm with a bootstrap approach to take draws from this posterior. For each draw, we bootstrap the data to simulate estimation uncertainty and then run the EM algorithm to find the mode of the posterior for the bootstrapped data, which gives us fundamental uncertainty too (see Honaker and King (2010) for details of the EMB algorithm).

alexwhitworth
  • 4,839
  • 5
  • 32
  • 59