Subspace clustering method for high dimensional data stream
一种适用于高维数据流的子空间聚类方法

Inspired by the FP algorithm used in mining frequent patterns and the idea used in CLIQUE which is a classical method for clustering static data,a new data structure named dense grid-tree(DG-tree for short) was proposed to record the synopsis of the data streams for clustering.Then the clustering problem was transformed to the problem of constructing a DG-tree and searching for dense grid cells in the DG-tree.With the help of DG-tree,the subspace containing clusters was found.Experimental results show that this method has good cluster quality.