Best approach for NaN labels

### Problem description Hi, I have more like a question than a problem, but I guess this is the right place. ### Feature description I would like to ask what is the best approach when I have NaN values in my dataset on the columns that will be used as labels. I do not really want to remove NaNs but I want to know how this will affect the MajorityVote and the Dawid-Skene classification and what's the best way to have them in the data set i.e. as np.nan? blank spaces? Thank you, Lydia ### Potential alternatives _No response_ ### Papers connected with feature _No response_ ### Additional information _No response_