A Framework for Uncertain Spatial Data Mining and Its Application to Environmental Geochemistry Quality Assessment

摘要

A framework for uncertain spatial data mining was proposed. In which, the uncertainties of spatial data themselves and spatial data mining are emphatically dealt with, including uncertainty simulation by Monte Carlo method, spatial autocorrelation matrix based on uncertain spatial data, discretization based on neighborhood EM algorithms, and quality assessment of results. Meanwhile, the experiments of environmental geochemistry quality assessment with uncertain and spatial data mining have been performed using the environmental geochemistry data gotten from Dexing, Jiangxi province in China.