论文部分内容阅读
传统的聚类方法能够自动实现类型的划分,但其只考虑空间对象的属性特征,忽略了空间对象之间位置的邻近性。在尺度空间理论基础上,利用多尺度空间聚类方法,同时兼顾空间对象间的空间相邻和属性相似这两个特征,使得在空间尺度由小变大的过程中,具有高度空间相互依赖关系的空间对象相互融合,得到不同空间尺度下的区域划分结果。并以福建省67个县市1990—2004年历年的人均GDP数据分析为例,进行了全省的经济区域划分应用试验。结果表明与实际的区域经济发展情况相符。该方法能够综合考虑空间位置邻近性和属性特征相似性的聚类要求,使得聚类结果更加合理,为制定区域经济持续发展战略提供依据。
The traditional clustering method can automatically classify the types, but it only considers the attribute features of the spatial objects and ignores the proximity of the spatial objects. On the basis of the theory of scale space, using the multi-scale spatial clustering method, taking into account the two characteristics of spatial adjacent and attribute similarity between spatial objects, making the spatial scale interdependent in a high degree of spatial scale from small to large, The spatial objects are integrated with each other to obtain the result of regional division under different spatial scales. Taking the per capita GDP data of 67 counties and cities in Fujian Province during the 1990-2004 calendar year as an example, this paper carried out the application test of the economic area of the province. The result shows that it is consistent with the actual regional economic development. This method can comprehensively consider the clustering requirements of spatial location proximity and attribute similarity, which makes the clustering result more reasonable and provides the basis for formulating the sustainable development strategy of regional economy.