Variable-wise imputation errors

In the previous exercise you have extracted the estimated imputation errors from missForest's output. This gave you two numbers:

  • the normalized root mean squared error (NRMSE) for all continuous variables;
  • the proportion of falsely classified entries (PFC) for all categorical variables.

However, it could well be that the imputation model performs great for one continuous variable and poor for another! To diagnose such cases, it is enough to tell missForest to produce variable-wise error estimates. This is done by setting the variablewise argument to TRUE.

The biopics data and missForest package have already been loaded for you, so let's take a closer look at the errors!

This exercise is part of the course

Handling Missing Data with Imputations in R

View Course

Hands-on interactive exercise

Have a go at this exercise by completing this sample code.

# Impute biopics data with missForest computing per-variable errors
imp_res <- ___(___, ___ = ___)