Study On Method And Application Of Land Use Spatial Data Mining

Land manager department of the government has obtained a lot of data by land use actuality investigate and urban cadastral survey construction, but how to get useful information is the important problem that must be resolved. The paper made study about mine data on land use data by resolving the problem of how to select tools, algorithm, and how to analyze the results. This paper is mainly divided into three greatly part of: the first part introduced basis, domestic and international research present condition, technical route etc.; the second part introduced the algorithm that was used in the paper; the third part is three examples.By making use of different software and algorithm, three examples drew some conclusions:(1) The paper uses SQL Server 2000 Analysis Manager and decision trees to mine data on Ji'nan cadastral information in 2000. The result showed that floor area ratio of Ji'nan is low. So use efficiency of urban land should be improved in future.(2) The paper uses gini coefficient to analyze differences of urban construction land. The result showed:①Gini coefficient of urban construction land based on one kind of land style in four areas is very small and its development tendency is different, but there's a huge difference among areas.②Gini coefficient of urban construction land in most of provinces descended slightly, and it explains gini coefficient of urban construction land tended to rationalization.③According to gini coefficient, we contrasted the result and GBJ 137-90. The gini coefficient was a useful way to analyze differences of urban construction land.(3) The paper tries to use Weka to analyze the structure of urban construction land use and gets some useful results. Some means will be used, containing many algorithms of associate, cluster and regression.①It has been get from associate that there is relativity in different kinds of land.②All cities have been divided into 5 types by cluster. And the spacial distributes characteristic of 5 types was analyzed.③It has been get from regression that there is relation in land structure and city size.(4) Forming technical workflows to resolve the problem of how to select tools, algorithm, and how to analyze the results.