滑坡预报的多元回归分析方法

The multivariate regression analysis method for landslide prediction

在回顾了滑坡预报的概念、类型及已有滑坡预报的几种主要模型的基础上,利用多元回归分析方法,根据最小二乘法原理,建立了非线性回归预报模型,提出滑坡破坏的时间为位移 时间曲线的拐点,此方法一般适合于临滑预报。结合实例,利用黄龙西村滑坡位移 时间监测资料,采用该模型进行滑坡时间预报,和其他滑坡预报模型预报的结果接近,说明该模型具有一定的合理性。

This paper firstly states the conception and types of landslide prediction,then introduces several main methods for landslide prediction.We use the method of multivariate regression analysis and the principle of least\|square procedure to build the predicting model of nonlinear correlation,and put forward that the slipping time is the point of inflexion on a displacement\|time curve.The method is fit for the prediction of approaching slippage.We use the model to deal with the monitoring displacement\|time data of Huanglongxi village landslide and predict the time of its slippage.The result is close to those of the other models. Therefore it is proved that the method of multivariate regression analysis is suitable for landslide prediction.