史前连续文化聚落的决策树分类挖掘研究--以郑州-洛阳地区为例  

Research on classification mining of decision tree for prehistoric continuous culture settlements

作　　者：毕硕本[1] 闾国年[2] 陈济民[3] 

机构地区：[1]南京信息工程大学计算机学院,南京210044 [2]南京师范大学地理科学学院,南京210097 [3]黄河勘测规划设计公司测绘信息工程院,郑州450045

出　　处：《测绘科学》2008年第2期161-164,共4页Science of Surveying and Mapping

基　　金：中国博士后基金项目(2005037743);江苏省博士后基金项目(0501019B);江苏省教育厅计划先导项目(05KJD170109)

摘　　要：利用C4.5决策树分类算法和GIS组件Super Map Objects,开发实现了基于连续文化序列的空间数据挖掘系统,运用此系统对郑州-洛阳地区史前四个连续文化时期的聚落进行了数据挖掘,提取了聚落遗址面积的分类规则,并根据这些分类规则和聚落的空间分布情况对四个文化时期的I级聚落进行对比分析,得出了I级聚落的空间分布特点,即I级聚落中的一些特大面积的聚落朝着中心聚落的方向发展。Using classification algorithm C4.5 of decision tree and GIS components SuperMa PObjects, the spatial data mining system based on continuous culture settlements is developed. The system is applied to the data mining for the prehistoric settlements of the four culture periods about Zhengzhou-Luoyang region. The area classification rules of the sites are excavated. The spatial distribution characteristic of level-Ⅰ settlements are gotten by contrasting and analyzing the classification rules and the settlement distribution of level-Ⅰ settlements. The characteristic is that the large area settlements of level-Ⅰ settlements are expanded to the central settlement.

关 键 词：空间数据挖掘 决策树分类 聚落考古 连续文化 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]