均質点集合を用いた農村集落の分類

Abstract

There exist many kinds of objects which can be regarded as a poin set. But if one calculates density of a point set he will find that sometimes it is not so easy, since it is difficult or even impossible to identify or draw the boundary of the set. Classification of rural settlements byconsidering density of housing is one of these problems.This study, regarding housing in a village as a point set and comparing their density with the “uniform point set” which is supposed to spread homogeneously, set up two quantitative classification indices, that is scatter index and scatter revision index, and then use them to classify villages by two steps.The bigger the scatter index, the more scattered the village would be. By using this scatter index, rural settlements are divided, at the first place, into four types. They are “mikyo”, “shukyo”, “sankyo”and “sokyo” in Japanese, indicating density, semi-density, semi-scatter and scatter respectively (more then four types, however, could be classified if necessary) . Then the scatter revision index is used to further classify those settlements whose scatter indices are too close to the standard value to be rationally dealt with. In this case a larger scatter revision index indicates a denser distribution of housing in village. Finally, it should be noted that the method developed here to classify rural settlements could also be applied to other similar classification problems which in somewaycould be changed into point set.