A Method of Multisource Geospatial Vector Data Mining Based on GIS
Under the interference of noise data,the accuracy and efficiency of data mining are both low.In order to improve the accuracy of data mining results,a multi-source geographic vector data mining algorithm based on GIS was proposed.First,the undecimated wavelet transform was used to perform multi-scale decomposition on the original multi-source geographic vector data,thus ensuring that noise and data could be effectively separated.Then,variational partial differential equations were used to smooth the residual noise in the data,thus avoiding the loss of detailed information in the data.Next,GIS was applied to the multi-source geographic vector data mining.And a data mining model was built.Meanwhile,the data was transformed according to the relationship between layers,thus dis-cretizing multi-source geographic vector data.Moreover,conceptual lattice and Hasse diagram were constructed based on formal contexts.Finally,non-redundant association rules were formed by the intension reduction set in the concep-tual lattice.Thus,the data mining task was completed.Experimental results show that the proposed method can effec-tively filter out the noise in the multi-source geographic vector data and obtain satisfactory data mining effects.
NETL
NSTL
万方数据
