相空间局域预测法在滑坡位移预测中的应用

Prediction of Landslide Displacement by Local-region method

根据滑坡系统的混沌特性,将适合于混沌时序预测的相空间局域近似法应用到滑坡位移预测中。针对零阶局域预测法利用信息效率不高的缺点,运用模糊数学中构造隶属函数的方法来选取预测相点的近邻点,减小了选取近邻点这一关键问题的人为随意性,而且预测结果也证明了这种方法的合理性。

According to the chaotic characteristic of landslide dynamic system, the local-region method which is suitable for chaotic time series may be applied to predicting the displacement in landslide. For the ordinary zero-rank local-region method can＇t take full advantage of data, a fuzzy theory is introduced in order to select the nearest neighbor points in phase space. This method can reduce the subjectivity in selecting nearest neighbor points to some extent and the calculated results also prove this method is correct.