基于灰色最小二乘支持向量机的边坡位移预测.pdfVIP

第31 卷第5 期 岩 土 力 学 Vol.31 No.5 2010 年5 月 Rock and Soil Mechanics May 2010 文章编号：1000－7598 (2010) 05 －1670－05 基于灰色最小二乘支持向量机的边坡位移预测 马文涛 （宁夏大学数学计算机学院，银川 750021 ） 摘 要：利用边坡实测位移序列预测边坡未来时间的位移，可以有效地判断边坡的稳定性。在分析了灰色预测方法和最小二 乘支持向量机各自的优缺点的基础上，提出了将二者相结合的一种新的预测模型——灰色最小二乘支持向量机预测模型。新 模型既发挥了灰色预测方法中“累加生成”的优点，弱化了原始序列中随机扰动因素的影响，增强了数据的规律性，又充分 利用了最小二乘支持向量机求解速度快、易于描述非线性关系的优良特性，避免了灰色预测方法及模型存在的理论缺陷。同 时，采用遗传算法进行了模型的参数优化，通过2 个工程实例说明灰色最小二乘支持向量机模型预测边坡位移的有效性，具 有较高的精度。 关 键 词：边坡位移；灰色模型；最小二乘支持向量机；遗传算法；时间序列 中图分类号：TU 471 文献标识码：A Forecasting slope displacements based on grey least square support vector machines MA Wen-tao (College of Mathematics Computer Engineering, Ningxia University, Yinchuan 750021, China) Abstract: Based on the displacement sequence of slope, the stability of slope could be judged effectively by forecasting the displacement of slope in the future. Through analyzing advantages and disadvantages of grey forecasting methods and least square support vector machines(LSSVM) respectively , a new forecasting model of grey least square support vector machine was proposed. The new model not only developed the advantages of accumulation generation of the grey forecasting method, weakened the effect of stochastic-disturbing factors in original sequence and strengthened the regularity of data, but also used the quickly solving speed and the excellent characteristics of least square support vector machines for nonlinear relationship and avoided the theoretical defects existing in the grey forecasting model. At the same time, the genetic algorithms were used to optimize the parameters of new