Improving Identification of Difficult Small Classes by Balancing Class Distribution.
Jorma LaurikkalaPublished in: AIME (2001)
Keyphrases
- class distribution
- highly imbalanced
- majority class
- class labels
- rare classes
- target class
- class imbalance
- training data
- minority class
- cost sensitive
- test set
- training samples
- concept drift
- training set
- misclassification costs
- small number
- highly skewed
- imbalanced data
- test data
- imbalanced datasets
- training examples
- cost sensitive learning
- classification error
- training set size
- small disjuncts
- imbalanced data sets
- data mining
- high dimensional data
- data streams
- reinforcement learning
- decision trees