make sure to clean the data by looking for duplicates or typos so no one community is treated as two or more this can make unstabilty..you can try multiple types of encoding but what I found the most usefull is target encoding and frequency encoding but you must be aware ofthe overfitting.
make sure to clean the data by looking for duplicates or typos so no one community is treated as two or more this can make unstabilty..you can try multiple types of encoding but what I found the most usefull is target encoding and frequency encoding but you must be aware ofthe overfitting.