Drift Lens: Real-time unsupervised Concept Drift detection by evaluating per-label embedding distributions.
Salvatore GrecoTania CerquitelliPublished in: ICDM (Workshops) (2021)
Keyphrases
- concept drift
- real time
- novelty detection
- change detection
- drift detection
- data streams
- streaming data
- drifting concepts
- non stationary
- data stream classification
- incremental clustering
- data stream mining
- classification algorithm
- data distribution
- detection algorithm
- evolving data streams
- batch learning
- remote sensing
- class labels
- sliding window
- image processing
- learning algorithm