Acceleration of Functional Cluster Extraction and Analysis of Cluster Affinity

In this paper, we address the problem of extracting functionally similar regions in urban streets regarded as spatial networks. To efficiently deal with several large-scale networks, we propose a fast extraction method of functionally similar regions using the lazy evaluation and pivot pruning techniques. In our experiments using the urban streets of 12 cities from all over the world, compared with a state-of-the-art method based only on the lazy evaluation technique, we show that our proposed method achieved a reasonably high acceleration performance. We also show that our method could extract major functional clusters as regions corresponding to downtown, suburban, and mountainous areas for all the 12 spatial networks used in our experiments, and each cluster for the same area had quite similar characteristics in terms of the relations among the other clusters.