The Data Mining Of The Ground Deformation's Monitoring

In recent years,because of the large-scale exploitation of the mining area,the mining area appeared surface movement and deformation.It serious impacted to the production's safety.Conduct ground deformation monitoring and prediction has become the focus of the production safety.In a large number of monitoring data implied knowledge and laws of the surface's deformation,such as the mining area's multi-related and the mining area's overall sedimentation,these laws are benefit to surface deformation.Therefore,the paper take the aquifer of the mining area and the mining area's settlement as an example,use various theories of the Spatial data mining to gain implied law of the surface deformation from the surface deformation monitoring data mining,and to predict the surface subsidence. Elements include:Firstly,we brief introduced the definition and Features,the architecture,the basic process,the availability of knowledge types,mining methods and direction of development of Spatial Data Mining.Secondly,established the GM(1,1) model and linear regression model which could be used for the mining of mining area's subsidence data,and improved the GM(1,1) model.The paper used the GM(1,1) model for short-term forecasts,with the least absolute value method instead of the least-squares method.And we take the subsidence monitor of the coal mine main well as the example,confirmed the validity of the improvement method,and achieved satisfactory results.Linear regression model was applied to predict the long-term surface subsidence.Thirdly,take the aquifer of the mining area and the mining area's settlement as an example,applied the gray relation to solve the deformation factor's correlation to gain the law of the deformation factor's correlation.Fourthly,applied the clustering based on the grey relational to classify the Different Period's surface deformation of the mining area and the reason's data mining.