基于偏最小二乘通径模型的武汉市岩溶塌陷危险性预测

Application of the PLS Path Model to Forecasting Karst Collapse Hazard in Wuhan City

摘要:

目前对岩溶地面塌陷的研究仍主要在宏观和定性的水平上。偏最小二乘通径模型采用一系列一元或多元线性回归的迭代求解,无需对观测变量做特定的概论分布假设,对样本点容量的要求也十分宽松,不存在模型不可识别的问题,是一种实用和有效的线性统计建模方法。本文在分析武汉市岩溶塌陷影响因素的基础上,基于GIS技术,采用偏最小二乘通径模型对武汉市岩溶塌陷危险性进行了预测。结果表明利用该模型可以取得较好的预测结果,在工程实际应用中有一定的借鉴意义。

Abstract:

The study of karst collapses is still in macroscopic and qualitative level right now. The PLS Path Model adopts the iteration solution of monogenesis or multivariant linear regression, does not need to hypothesize the probability distribution of observation variables. In this model the requirement of sample point capacitance is very loose, and the problem that the model can not be recognized does not exist, so it is an effective linearity statistics modeling method. In this paper, based on the analysis of the influence factors for karst collapses in Wuhan and GIS technique, the PLS Path Modle is used to forecast the hazard potentiality of karst collapse in Wuhan city with a good results. It is found that the application of this model could get good effictiveness and this model should be adopted widely.