Adaptive Method for Handling Local Concept Drift of Data Streams Classification
一种自适应局部概念漂移的数据流分类算法

Based on the DB2 method,an adaptive method called IncreDB2 is proposed to detect and adapt to local concept drift continuously in data stream classification.This method dynamically maintains a hierarchical classification tree.When local concept drift is detected,IncreDB2 only updates the nodes that affected by this drift rather than rebuilds a new classification tree from scratch,which means that it has better time efficiency.Experimental results demonstrated the validity and efficiency of this method.