郑洛地区龙山文化遗址预测模型

Predictive model of archaeological sites of Longshan Culture in Zhengzhou-Luoyang area

本文以郑洛地区龙山文化时期的聚落遗址点为研究对象,借助地理信息系统软件的空间分析功能,获得遗址点和非遗址点的相关环境变量,采用二分变量逻辑回归模型进行分析,建立郑洛地区龙山文化遗址预测模型,并生成遗址点分布概率图。利用Kvamme增益统计进行模型验证,结果表明:遗址预测模型精度良好,能够揭露遗址的空间分布规律,为田野考古提供参考。

This paper took the settlement sites of Longshan Culture period as the research object, obtained the relevant environ- ment variables by spatial analysis in C, IS. Then binary logistic regression was used to create a settlement site prediction model to define areas of high, moderate and low potential for settlement sites of Longshan Culture period. Then the predictive model and associated probability map were produced for understanding the spatial distribution patterns of Longshan culture settlement sites in Zhengzhou- Luoyang area. Finally the model was validated by Kvamme＇ s Gain Statistic, and the results suggested that the predictive model could guide future field surveys to detect unknown sites in a more cost-efficient manner through its high precision.