Reasearch Of GIS-Based Spatial Data Mining About Settlement Site Groups In Zhengzhou-Luoyang Region

This paper researches the distribution and the historical evolution of settlement groups in Zhengzhou-Luoyang region, from the three elements of the spatial data including property, space and time. It utilizes different spatial data mining methods by means of the SQL Server database and GIS software.The research results of this paper are as follows:(1) The spatial data of the settlement groups are analyzed by the spatial data mining methods, which are decision tree classification, spatial analysis and spatio-temporal data mining and so on. The research makes spatial data mining methods deeper and the problem about explosion of data but lack of knowledge in Zhengzhou-Luoyang region’s settlement archaeology field is solved to some extent.(2) The properties of spatial data of the settlement groups are analyzed by the decision tree classification method. To avoid the multi-valued tendency of the ID3algorithm and improve the efficiency of the ID3algorithm generating decision tree, it puts forward the improved algorithm, which is named SA_ID3. SA_ID3algorithm is proposed by introducing the user interest and simplifying the process of ID3algorithm. The SAID3algorithm is applied to the classification mining on area property of settlement site groups about four continuous culture periods and the classification rules are extracted to prove the effectiveness of the SA_ID3algorithm. Therefore the spatial data mining on data mining applications is further expanded and the implementation efficiency of the existing data mining algorithms of the spatial data is improved.(3) In support of GIS, the space elements of the settlement groups’spatial data are analyzed by the spatial analysis method. The relationship of the settlement group distribution with the five environmental factors including the elevation, slope, aspect, distance from the water and the mountain, is analyzed by using the elevation, slope and aspect analysis, buffer analysis and other spatial analysis methods. On this basis, the settlement group object index model is established and the suitability index evaluation system is constructed. The multi-objective optimization results of the weighted analysis are compared with specific circumstances of the distribution of site groups to assess the rationality and the distribution reasons. The subject investigated and application range of spatial data mining on GIS archaeology applications is further expanded.(4) The time element of spatial data of settlement groups is analyzed by the spatio-temporal data mining method. The spatio-temporal data model of settlement groups is established based on the object-oriented. The center points and the areas of sub-settlement groups are computed by the center point and area algorithms. Then it computes the evolution speed of the center points, the site density and the integrated rating. Building data management system of the sub-settlement object spatio-temporal data model, the center point is used to flag the sub-settlement group. The spatio-temporal evolution of sub-settlement groups is analyzed from the evolution trajectory and speed of the sub-settlement group center point combined with the area, site number, and site density and integrated rating in various periods of sub-settlement groups. The dynamic characteristics over time of the spatial distribution of the sub-settlement group are clearly discussed and the field of spatial data mining research is extended.