I’m working on ML project. I have a df with some missing values that I don’t want to impute. Is there a way to numerically represent these NA values? Like a number that R is able to interpret as a missing value.

I’ve read that imputing NA values with -9999 shoul work work. Is it true? Or are there better ways to do it?

To be more specific, I then want to train some classification models for these observations. Would this numerical representation alter my results?