Spatial Data Mining For Huairen County Land Use Database

Spatial Data Mining(SDM) is a technology which can extract interested and hidden knowledge and spatial relationships. In recent years, with the implementation of the second national land survey, more detailed land survey data has been collected and a lot of information implied in land status of the database survey. Regular information can be extracted from existing database by technology of spatial data mining, which would increase using efficiency of actual data and expand application areas in data, also provide technology support in land use and management. Land use database in Huairen County was established on the basis of the project of "the second land survey of Shuozhou city Huairen county in Shanxi province". Combined with related SDM technology, methods and processes of spatial data mining in Huairen county land use database were studied, and the programming examples of SDM were designed and tested. The study included:(1) Based on industry data collection in the land survey of Huairen county in Shanxi province, the HuaiRen county land use database was built, to realize integrated management of data.(2) A new implementation scheme relates with spatial data mining in land use database was proposed, which used ArcGIS software, spatial data engine of ArcSDE and database of SQL Server2005 to implement the management of spatial data and spatial analysis, further implement the relevant spatial data mining tasks in Microsoft SQL Server2005 Analysis Services (referred to as SSAS).(3) Based on the needs of SDM data processing, the processes of spatial modeling were studied in GeoProcessing of ArcGIS to realize batch data processing modeling; The spatial data relationships were expressed in the form of property, using the function of abundant space analysis to extract the attribute data of space adjacent polygons in ESRI's ArcEngine as the second development component.(4) Using Microsoft clustering algorithm, the distribution around the land class was found under the influence of the factor of nearly round, and the mining results were analyzed and verified.(5) Using Microsoft Association rules algorithm to discover the relationships between the slope and the second land classes as well as the implied relationships between the land class and its adjacent land class, A series of practical and reasonable associated rules were obtained through the training model and setting different parameters. The models were proved accurate through practice.