A high-dimensional data stream clustering algorithm based on damped window and pruning list tree

In order to effectively reduce the memory consumption, a synopsis data structure, PL-Tree, is proposed, which can store the summary information of data streams and help to quickly output the clustering results when any clustering is requested at any time. Then, PLStream, an efficient high-dimensional data stream clustering algorithm based on PL-Tree and damped window is presented. Simulation and comparison experiments demonstrate that compared with the classic CELL TREE algorithm, PLStream has better performance in execution efficiency, spatial scalability and clustering effect.