I’m trying to create a custom error function to a cost sensitive, sample by sample, binary classifier problem
Consider a dataframe with X features y vector of 0/1, error_