Highdimensional data have the characteristics of sparsityand vulnerability to dimension disaster

which makes it is difficult to ensurethe precision and efficiency of high dimensional data clustering Therefore themethod of subspace clustering is adopted to reduce the impact of sparsity anddimension disaster on the clustering results Firstly

random sampling isadopted to select the dimension which is suitable for clustering from highdimensionaldata to generate subspace

and the hoeffding bound is combined to ensure thevalidity of sampling results Secondly

by using the adjacency of the grid

mixed grids are generated in the subspace

which can guarantee the integrity ofdata and improve the density of the subspace Finally

according to thesimilarity and dissimilarity of subspace

the dimension pruning is carried outto improve the subspace density again The algorithm can achieve better resultson UCI data set

and it has better performance in scalability and antinoiseability