Hi everyone,
Working on a large dataset from our institution. It came to our attention that we will need to account for missing data on some variables. When you use multiple imputation, is there any particular issue you will need to consider? I will review some of literature as well, but would appreciate having some quick tips here.
Sincerely,
Sangil Lee
You could use causal methods to impute missing. For e.g. generate binary var as miss/available and estimate propesity score and then weighting by inverse if ps.
Also, this method could be help full in evaluting randomness of missing.
The following video on data screening might be helpful: https://www.youtube.com/watch?v=iWrQ-SgVy-0 .
Good luck with your research!
Il est important d'identifier les données manquantes dans un jeu de données avant d'appliquer un algorithme.
Donnees manquante (Si il existe une feature pour laquelle sa valeur n'est pas rendeignee)
These are some excellent responses and I appreciate your willingness to help! SL
- Badges
- Science topic