The Role of Domain Expert Knowledge in the Process of Spatial Data Mining
Lan Rong-qin · Journal of Institute of Surveying and Mapping · 2004
Spatial data mining technique has been widely used in many areas, especially in intelligent GIS and spatial decision support system. Data mining has been often described as a process of automatic discovery of new knowledge from large amounts of data. But the role of the domain expert knowledge in this process and the forms it can take is an issue that has been omitted so far. Domain expert knowledge, on the one hand, guides the practice of spatial data mining, and on the other hand, continuously be renewed or replaced by the learned knowledge. That is to say, the spatial data mining process is a continuous interaction between domain knowledge discovered through the use of data mining algorithms. In view of the use of domain expert knowledge in the process of spatial data mining, this paper proposed that the domain expert knowledge should be classified as three categories, that is, Attribute Partition Rule, Concept Hierarchical Trees and Constraints. The role of domain expert knowledge and domain experts in the stages of data mining are discussed in detailed. Also, presentation of domain expert knowledge is an issue to which thw importance should be attached. A presentation model is given in the final section of this paper.