数据流集成分类算法综述

urrently, the trend of data stream classification algorithms is to ensemble classification algorithms. Because the ensemble algorithm provides better performance and more outstanding performance than the single classification algorithm. At the same time, it is easy to deploy in practical applications in the real world, has rapid adaptability and recovery to concept drift, and has the best classification performance in the processing of class imbalance problems. Based on the outstanding features and performance of the above ensemble classification algorithm, it has won extensive research by scholars at home and abroad. This paper introduces the ensemble classification algorithm at home and abroad in detail. The two parts of the ensemble classification algorithm (base classifier combination and dynamic update ensemble model) are reviewed in detail, and the advantages and disadvantages of different integration algorithms, comparison algorithm and experimental data set are clearly distinguished. The paper proposed further research Directions and considerations.