Divide-and-conquer algorithm for creating neighborhood graph for clustering

K-nearest neighbor graph has been used for reducing the number of distance calculations in PNN-based clustering. The bottleneck of the approach is the creation of the graph. In this paper, we develop a fast divide-and-conquer method for graph creation based on the algorithm previously used in the closest pair problem. The proposed algorithm is then applied to agglomerative clustering, in which it outperforms previous projection-based algorithm for high dimensional spatial data sets.