A Novel Neural Ensemble Architecture for On-the-Fly Classification of Evolving Text Streams

AdaNEN is a neural ensemble architecture for classifying text streams whose underlying distribution shifts over time — learning on the fly, in a single pass, without retraining from scratch.

It improves classification accuracy by up to 8.8% across 13 benchmark datasets. This work grew out of my M.Sc. thesis at Bilkent.

← All publications