A Kind of Clustering Algorithm Based on N-Tree

Due to the problem of the excessive time complexity in multi-dimension data-processing based on density algorithm, this paper proposes a clustering algorithm based on n-tree spatial index structure, which can realize the quick judgment of adjacent units of any unit through the reasonable encoding to the segmented units. By adopting the pruning in n-tree generating process, this algorithm can effectively save the cost of CPU and I/O. Compared with DBSCAN algorithm by different dimensions experiments, this algorithm can largely increase the capacity of processing massive, high-dimension set.