Active Feature Acquistion for Opinion Stream Classification under Drift
Proceedings of the Workshop on Interactive Adaptive Learning (IAL 2019), page 108--111. CEUR Workshop, (2019)Abstract
Active stream learning is frequently used to acquire labelsfor instances and less frequently to determine which features should beconsidered as the stream evolves. We introduce a framework for activefeature selection, intended to adapt the feature space of a polarity learnerover a stream of opinionated documents. We report on the first results ofour framework on substreams of reviews on different product categories.