When imputing a missing value (in my case using MICE) - should I use all the variables in the dataset or should I use only the variable which correlate most with the missing values I want to impute?

Thank you!

  • $\begingroup$ Can you expand to give more details and also clarify what you mean by "correlate most with the missing values"? $\endgroup$ – mdewey Jul 12 '17 at 10:32
  • $\begingroup$ I mean those variables who have low Pearson R or low score in a heatmap. In my case (working with a dataset of train passengers) I have missing values in "Age" variable, and I know that "Ticket Number" each passenger got is a random number, thus have nothing to do with age, but I fear the machine can take it as a correlation. $\endgroup$ – Riddle-Master Jul 12 '17 at 12:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.