Combining Diverse Meta-Features to Accurately Identify Recurring Concept Drift in Data Streams.
Ben HalsteadYun Sing KohPatricia RiddleMykola PechenizkiyAlbert BifetPublished in: ACM Trans. Knowl. Discov. Data (2023)
Keyphrases
- concept drift
- data streams
- data stream classification
- streaming data
- drifting concepts
- data stream mining
- non stationary
- ensemble classifier
- change detection
- stream data
- incremental clustering
- drift detection
- evolving data streams
- data distribution
- sliding window
- classification algorithm
- mining data streams
- feature vectors
- concept descriptions
- batch learning
- computer vision
- outlier detection
- classification accuracy
- feature extraction
- image processing
- data sets
- stream processing
- feature set
- prior knowledge
- feature space
- noisy data streams