Editorial

摘要

This issue opens with a path-breaking paper by Murtagh, whoshows that in large spaces, as well as in sparse spaces, the distances between data points tend to become ultrametric. In these circumstances, there is no “curse of dimensionality”, and it turns out from the benchmark studies given that clusters in the data become rather easily recognizable in the distribution of the distances (from which also the number of clusters can be determined).