Spatial Clustering-based Investment Decision Support System

Due to spatial data being more and more important during the process of discovering implicit knowledge and reconstructing nature, spatial data mining has been an important research aspect in data mining fields. The object of spatial data mining is to support the decision process by discovering useful knowledge from the abundant and complex spatial data sets.Spatial data owns much character which is different from that owned by ordinary data and the essential difference between them is spatial attribute. According to the development of spatial data mining and the hotspot of study and based on lucubrating spatial attribute, this paper proposes a method to pretreat spatial data by data pretreatment to transform it into proper form on which the method data mining can be used. The conception hierarchies and generalization technique are mainly used. Generalization technique is used after description of complicated relation between spatial objects such as topological relation, orientation relation and so on which are abstracted to the higher hierarchy. In addition, spatial attribute and non-spatial attribute are both pretreated.This paper develops a system of investment decision making which overcomes the shortcoming of actual system based on non-spatial data (attribute data of customers), ill-considered or neglecting spatial information. The system is proved that the method which transforms the spatial data into proper form by data pretreatment is valid. Finally, an experiment is done which is proved that the system can support decision-making.