关键词: 概念漂移，相似性，集成学习，数据流分类，加权多数投票

Abstract: Classification of data streams has become one of hot research spots, and a new similarity-based dynamic en- semble algorithm was presented to deal with two critical problems, namely, concept drift and noise. Because adjacent data in data stream has the same concept with more probability, the new sub-classifier stands for the coming concept Based on it, ensemble classifier is got by similarity weighted majority voting, and sulrclassificr with worst performance is dele- ted for suiting for concept drift and noise. Experiment result on simulation data set shows the algorithm is best than other schema in classification accuracy and anti-noise.

Key words: Concept drift, Similarity, Ensemble learning, Data stream classification, Weighted majority voting