Variable Weight Combination Method In Landslide Prediction

Landslide disaster is one part of natural environment, it is widespread on the global scale and is a serious hazard next only to earthquake. It often destroy buildings, lead to traffic jam, cause casualties, did serious harm to the ecological environment and engineering construction, also bring about a lot of losses. It is impossible in our country to have enough economic and technical force to thoroughly control the risky landslide. So using new methods, aiming at raising the accuracy of prediction, taking steps to prevent and reduce natural disasters, reducing the loss caused by this kinds of disaster to the lowest degree, in the research of landslide is valuable.Firstly, a series of problems about the failure time prediction are discussed in this paper. The problems mentioned above include the theory and method to divide the failure time, the method and principle to select and process parameters used by the failure time prediction, the step to establish the criterions of prediction, the principle to classify and select the prediction models. At the same time, this paper elaborates on some practical forecasting model in landslide. According to its applicability, we divide it into two phases of long term and short term,and the long term include Time series forecasting model, Nonlinear dynamics model and Grey-movement forecasting modes of uniformly factal. The short term include Verhulst model, Cooperative forecasting model, Exponential smoothing method and Moving average method.Then, lead into changeable weight combination forecasting model which based on multiple forecasting methods. It can comprehensive use the advantage of various methods and improve the precision accuracy of forecasting. And also, it can enhance practicablity of the model.At the end, the paper use the changeable weight combination forecasting method which based on Grey model, Verhulst model and Cooperative model to forecast the displacement of Xintan landslide in Zigui country, Hubei province. And it gain preferable effect.