To avoid training on imbalanced datasets, we usually sample data to a balance manner and ensure that every fold in cv contains all kinds of labels.
I'd like to find API like this, but I find stratified sampling and cross validation independently. Do we have methods to combine them together? Or how to use them collectively?
To avoid training on imbalanced datasets, we usually sample data to a balance manner and ensure that every fold in cv contains all kinds of labels. I'd like to find API like this, but I find stratified sampling and cross validation independently. Do we have methods to combine them together? Or how to use them collectively?