Impute missing variables for out-of-sample individuals using pre-trained FAMD model

Skip to first unread message

Dylan Mann-Krzisnik

Jul 29, 2021, 10:57:07 AM7/29/21
to FactoMineR users
Hello FactoMineRs,

I have trained an FAMD model using data (which had required imputation using imputeFAMD). I then use this model to predict the FAMD projections of new out-of-sample individuals.

What would be the best way for imputing data for new individuals with missing variables? As a constraint, the FAMD model should not be retrained, meaning that I would not simply be adding the new out-of-sample individuals as supplementary individuals at the time of training FAMD. In other words, the sequence would be:

impute sample data >> train FAMD model >> impute data of new individual >> predict FAMD projection of new individual

Perhaps my question requires clarification, which I'd be glad to provide.


Reply all
Reply to author
0 new messages