基于SQL Server Analysis Service（SSAS）的土地地类关系空间数据挖掘

摘要

基于SQL Server Analysis Services （简称SSAS）提供的Microsoft 关联规则挖掘算法和事务数据挖掘功能，通过利用ArcGIS软件、空间数据库引擎ArcSDE和数据库SQL Server软件，提出了一种新的土地地类关系挖掘实现方案。首先结合空间数据挖掘（ Spatial Data Mining ，SDM）相关技术方法，以土地利用数据库为基础，实现空间数据提取；然后通过空间关联操作将空间信息转化为事务，最后在SSAS中创建多维数据集，完成相关数据挖掘任务。基于某市实例土地利用数据库，采用该方法探测相邻地类间的隐含关系，通过建立相邻地类图斑空间关联规则挖掘模型，设置不同的参数，得到了一系列比较实用合理的关联规则，通过实践证明了这种方案的有效性。%Based on Mining algorithm of Microsoft associated rule and transaction data mining functions provided by SQL Server Analysis Services ( SSAS) ,a new implementation scheme for the land class relationships mining is proposed by using SQL Server,ArcGIS software,spatial data engine of ArcSDE .Firstly,Combined with related Spatial Data Mining ( SDM) technologies ,spatial data was extracted on the basis of the land use database .Secondly , the spatial data was transformed into transaction data by spatial associated rule .Finally,multi-dimension data set was constructed in SSAS and the relevant spatial data mining tasks were implemented .As a case study ,the implement scheme was applied to dis-cover the implied relationships between the adjacent land classes for the land use database of a county .Through training models and setting different parameters ,a series of practical and reasonable associated rules were obtained .It shows that the scheme is feasible and efficient in practice .