This thesis studies the diversity issue of classification ensembles for class imbalance learning problems. Class imbalance learning refers to learning from imbalanced data sets, in which some classes of examples (minority) are highly under-represented comparing to other classes …
"2" Result
Sort By: