Data Stream Classification

Concept Evolution

Phenomenon distinct from concept drift where new classes emerge in the data stream over time. Detecting concept evolution is critical for maintaining the relevance of classification models in environments where labels can evolve.

← Indietro