Knowledge acquisition model of map generalization based on granular computing
Ying Song, Dongmei Yu, Chen Shen · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2009
The knowledge of automated map generalization mainly derives from map specifications, experience of experts and spatial data. The representation, acquisition and reasoning of knowledge for cartographic generalization have been widely recognized as difficulty. In an example with results of cartographic generalization, it is certain that some special knowledge consist in spatial data, which represent spatial relationship of geographical objects. The meaning of acquiring the knowledge hided in data lies in that other unknown data can be inferred by the knowledge. In this paper, a model of knowledge representation and acquisition for automated map generalization based on granular computing is first proposed. Then, the conceptions concerning knowledge granule and its structure are defined, and intrinsic mechanism and method of knowledge acquisition are further discussed. Lastly, the model and method mentioned above are illustrated through a case study. The conclusion is that knowledge acquisition lies on the dependence degree of decision-making attributes for condition attributes. Each attribute has a different effect on the result of cartographic generalization. The decision-making rules knowledge acquired by difference of dependence degree is just the representation condition of cartographic objects, by which other unknown data with similar distribution characters can be inferred.