Hot-deck tricks & tips II: sorting by correlated variables
Another trick that can boost the performance of hot-deck imputation is sorting the data by variables correlated to the one we want to impute.
For instance, in all the margin plots you have been drawing recently, you have seen that air temperature is strongly correlated with sea surface temperature, which makes a lot of sense. You can exploit this knowledge to improve your hot-deck imputation. If you first order the data by sea_surface_temp, then every imputed air_temp value will come from a donor with a similar sea_surface_temp. Let's see how this will work!
This exercise is part of the course
Handling Missing Data with Imputations in R
Exercise instructions
- Hot-deck-impute the missing values in
air_tempin thetaodata, ordering bysea_surface_tempand assign the result totao_imp. - Create a margin plot of
air_tempvssea_surface_temp.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Hot-deck-impute air_temp in tao ordering by sea_surface_temp
tao_imp <- ___(___, ___ = ___, ___ = ___)
# Draw a margin plot of air_temp vs sea_surface_temp
tao_imp %>%
select(air_temp, sea_surface_temp, air_temp_imp) %>%
___(___ = ___)