| The data that the research institutes of settlement archaeology need to process are multifarious and disorderly. As one of the constant archaeological questions for discussion, the space-time issue is a more complex concept. It is greatly associated with the human culture mode, so it is very difficult to reorganize and process the complicated space-time data with the traditional archaeological methods of spatial analysis. Spatial data mining is to draw the concealed knowledge and spatial relations or the meaningful features and modes from the spatial database. This technology has particular advantages in comprehending spatial data and gaining the internal relations between the spatial and unspatial data.Taking the disadvantages of the traditional archaeological analytic methods and the advantages of the spatial data mining technology into consideration, I decide introduce the spatial data mining technology to the settlement archaeological research of space-time issue. Mining the data in the settlements of the four continuous culture periods in Zhengzhou and Luoyang area, I aim at discovering the settlement distribution law and the settlement space-time evolvement law, both of which are hidden in the large amount of data.This thesis expands some theories and methodologies closely related to the spatial data mining, such as, decision tree classification, spatial clustering and spatial analysis etc. I also analyse the basic procedure of spatial data mining and study the concrete algorithm in a deepgoing way .By making use of SuperMap Objects, I develop the spatial data system, which is based on the continuous culture sequences. After mining the data of settlements of the four periods in this area, I extract the spatial classification and clustering rules. I contrast the rules and discover the distribution regularity and space-time evolvement regularity of the prehistoric settlements in Zhenzhou and Luoyang area.(1) By using the decision tree classification, I extract the classification rules of the settlement site area. With these rules and the distribution situation of the settlements, I contrast the first-class settlements of these four cultural periods and find the features of the settlement distribution and the regular pattern of the space-time evolvement. I've got a conclusion that some of the outsize settlements of the first-class develop into the central settlements.(2) After analyzing with K-means clustering, I draw out the spatial clustering rules of the settlement site. During the YangShao and LongShan periods,the settlements ,which are near to the centre, are mainly the first-class and second-class. These settlements are similar to the centre in the model of the Central Location theory. They possess several features of the central settlements, which prepare themselves for becoming the capitals of dynasties ahead.(3) Using the spatial-analysis, I discover that these settlement sites are located about 4 kilometres to the water system. This illustrates that the water resource influenced the people's choices for residence in prehistoric times. People tend to choose the area which is near to the watersystem in order to gain abundant water resource and convenient traffic.(4) By analyzing the elevation of the sites of the settlement groups in different culture periods, I find that the elevation of the most settlement sites is among 100 to 300 meters,and the number of sites whose elevation is lower than 100 meters is less than 10%. What is more,the... |
