An active learning system for mining time-changing data streams

Abstract

Mining time-changing data streams is of great interest. The fundamental problems are how to effectively identify the significant changes and organize new training data to adjust the outdated model. In this paper, we propose an active learning system