Thanks for reply. Your observation is consistent with my chi-square test, which is Good group people tend to have more “debt consolidation “ Loan reason. Because both of them are relationships between “Loan reason “and “BAD“.. But when we observe the credit model, best group doesn’t seem to have more “DebtCondolidation” than worst group. That’s what I’m confused.
If the data is very skewed, for example, sometimes there are many same value credit scores for different customers, should we avoid to put same score to different category?