Lift-Per-Drift: An Evaluation Metric for Classification Frameworks with Concept Drift Detection.
Robert AndersonYun Sing KohGillian DobbiePublished in: Australasian Conference on Artificial Intelligence (2018)
Keyphrases
- drift detection
- evaluation metrics
- concept drift
- classification accuracy
- feature set
- feature extraction
- text classification
- data streams
- supervised learning
- support vector
- benchmark datasets
- information retrieval
- training set
- feature vectors
- active learning
- machine learning
- document collections
- precision and recall
- decision trees