期刊名称：ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
印刷版ISSN：2194-9042
电子版ISSN：2194-9050
出版年度：2002
卷号：XXXIV Part 2
页码：401-404
出版社：Copernicus Publications
摘要：This paper emphatically discussed the theory of Formal Concept Analysis, and discussed the basic theory of spatial data-mining. According to the thought of Formal Concept Analysis theory, this paper put forward some new thoughts of spatial data-mining. Formal Concept Analysis (FCA), is a kind of very efficient tool for data analysis. It is a set-theoretic model that mathematically formulates the human understanding of concepts, and investigates the possible concepts in a given domain. Each node of concept lattice is a formal concept which includes two sections: extent, intent. Through Hasse graph, concept lattice can lively, simply and clearly embody the generalization relationship and characterization relationship among these concepts. The procedure to form concept from datasets materially is a kind of procedure of concept clustering. Formal Concept Analysis not only can find out the hierarchical clustering, but also can find out a good description about concept. Concept lattice can be used in many machine learning tasks, its main shortcoming is high complicity. To control the increase of node is very necessary. Spatial Data-mining, means extracting interesting spatial modes and characters, the universal connection between spatial data and non-spatial data from spatial databases, and other universal data characters which implied in the spatial databases. Spatial databases implies large knowledge about the surface features and their relationships among each other, These knowledge have very important values to remote sensing image understanding, GIS spatial analysis, spatial decision support etc.. It is a very good method to apply Formal Concept Analysis in spatial data-mining tasks.
