A Self-Adaptive Framework for Concept Drift in Streaming Classification: Shadow-Model Pre-Training, Reliability-Weighted Detection and Scored Strategy Selection
AI-generated summary Researchers describe a streaming framework that trains a shadow model when detectors enter a warning state and combines alarms from ADWIN, DDM, and PageHinkley using online weights. The framework achieved a mean accuracy of 90.82 percent in experiments comparing its performance against seven baselines across multiple data streams.