Quantifying trends accurately despite classifier error and class imbalance.
George FormanPublished in: KDD (2006)
Keyphrases
- class imbalance
- majority class
- binary classification problems
- cost sensitive learning
- class distribution
- minority class
- active learning
- cost sensitive
- training data
- highly skewed
- feature selection
- imbalanced datasets
- training set
- high dimensionality
- sampling methods
- small disjuncts
- concept drift
- training samples
- imbalanced data
- training examples
- generalization error
- error rate
- class labels
- classification error
- weak classifiers
- training instances
- decision trees
- learning algorithm
- misclassification costs
- test data
- support vector machine
- feature space
- pattern recognition
- machine learning
- data sets
- classification algorithm
- supervised learning
- small number
- multi class
- data analysis