（南京工业大学 交通学院，江苏南京 210009）

[摘要]城市土地抗震适宜性评价可以通过土地发生地震灾害的概率大小来划分等级，而土地地震灾害的发生与否，是典型的二分类变量问题，运用Logistic回归模型处理这类（0，1）区间的概率问题，可以得到较理想的拟合效果，为城市土地抗震适宜性评价提供较合理的量化指标。而进行Logistic回归分析需要较大的样本数据，原始资料数据显然不能满足分析要求，这就需要利用GIS的空间分析功能进行空间插值，得到足够的样本数据，运用SPSS数值运算软件，进行二分Logistic回归分析。本文通过得到城市土地抗震适宜性的量化评价方法，以南京市为例，分析南京市具体地质情况，得到评价体系，并对评价指标赋值，通过运用Logistic模型得到评价指标权重和城市土地地震灾害概率公式，利用计算得到的地震灾害概率，实现定量化评价土地抗震适宜性的目的。

[关键词]抗震适宜性;Logistic回归模型;GIS;SPSS

中图分类号：TU435      文献标识码：A      文章编号：1002-848X(2013)S1-1479-04   

作者简介：袁媛，硕士，Email：358751744@qq.com。

Research on Evaluation Method for Urban Land Aseismic Suitability Based on the Logistic Model——With Nanjing as an Example

YUAN Yuan,CHEN Xinmin,YUAN Zongpan

(College of Transportation, Nanjing University of Technology,Nanjing Jiangsu 210009 China)

Abstract: Evaluation on urban land seismic suitability can be graded by the probability of earthquake disaster. Earthquake disaster of land occur or not is a typical problem of binary variables. Using Logistic regression model to deal with this kind of (0,1) inter probability problems can get ideal fitting effect. It can provide more reasonable quantitative indicators about evaluation on urban land aseismic suitability. Logistic regression analysis requires a large sample data. Logistic regression analysis requires a large sample data. The original data obviously can not meet the requirements for analysis, which requires the use of spatial analysis capabilities of GIS spatial interpolation to obtain a sufficient data. Then, the use of SPSS numerical computing software is to analyze binary Logistic regression model. In this paper, the quantitative evaluation method of urban land seismic suitability can be used with Nanjing as a example. Firstly, the evaluation system and evaluation index assignment of Nanjing can be get through studying the specific geological conditions of Nanjing. Secondly, the weight of evaluation indicators and the formula of the probability of the earthquake disaster on urban land can be obtained by using Logistic model. Finally, using the calculated probability of earthquake disaster was used to achieve the purpose of quantitative evaluation of land seismic suitability.

Keywords: seismic suitability; Logistic regression model; GIS; SPSS