Cloud Segmentation of Spatial Concept Hierarchy Based on Data Field

As spatial data mining involves geographic space and attribute space, both of which have a high relevance to each other, how to synthetically extract abstract concept of the two spaces to a higher hierarchy is becoming a new hotspot. Reviewing former approaches, it is possible to discover that the relationship between these two spaces is simply exhibited by weight in these approaches. Furthermore, in these approaches, observed samples are regarded as single objects, and their relationship as well as influence to parent space is ignored. Aiming at these issues, based on prevenient pan-concept-tree arithmetic, this paper integrates data field and synthesized cloud model with generation of pan-concept-tree to put forward the cloud segmentation of spatial concept hierarchy based on data field. By experiment and comparison, it is testified that this method has the capability to achieve the process of concept climb effectively and accurately.