Abstract: Site selection is one of the most crucial decisions made by any business. The existing approaches for site selection（commonly used by economists） are subjective, not scalable. Spatial co-location patterns represent the subsets of spatial features whose instances are frequently located together in geographic space. Thus, a novel method for site selection based on co-locations is proposed using the “symbiotic” characteristics among spatial features in co-locations. This method uses ontologies to classify the input data effectively, and then performs a co-location mining on the certain POI given by the user based on ontologies, the result co-location patterns are then used to conduct the method to give the final site recommendation in a certain location specified by the user, meanwhile, several strategies are performed in data pre-processing in order to improve the effectiveness of the proposed method. To assess the effectiveness of the presented method, it is evaluated on a real data （collected from Points of Interests in Beijing）. Evaluation results show that the proposed method has high accuracy ratio, which means high reliability in site selection.

Key words: spatial data mining, co-location pattern, ontology, site selection

摘要： 选址问题是任何一个商业机构都要面临的重大决策问题之一，它受多种因素制约，比如社会经济学、地质学、生态学以及决策者的特定需求等。现有的选址方法（通常被经济学家采用）大多利用主观评价，可扩展性差。空间co-location模式挖掘是空间数据挖掘的一个重要研究方向。一个频繁co-location模式是一组空间特征的子集，它们的实例在空间中频繁关联。利用co-location模式的这种特征间“共存”关系，提出了一种基于co-location模式的地址选择算法，该算法基于本体描述空间数据的分类信息，并在本体的指导下对用户感兴趣的兴趣点（Point of Interest）进行关键co-location模式挖掘，同时针对实际情况对数据进行了预处理以增加算法的有效性。在真实数据集（北京市的兴趣点数据）上的评估实验显示该算法具有较高的准确率，选择的地址具有高可靠性。

关键词: 空间模式挖掘, co-location模式, 本体, 地址选择