基于灰狼支持向量机的非等时距滑坡位移预测

以三峡库区白水河滑坡为例，针对滑坡位移监测数据的非等距性和复杂性，结合非等距时间序列分析法、灰狼优化算法（GWO）和支持向量回归机（SVR）模型，提出新型非等距位移时序预测模型.利用自然三次样条插值法对滑坡位移数据进行等时距处理，基于时间序列分析理论将位移数据中的趋势成分和周期成分剥离，采用基于稳健最小二乘法的三次多项式拟合和GWO-SVR耦合模型分别对这两者进行预测，利用时间序列加法模型得到滑坡累计位移的预测值.研究表明，基于灰狼支持向量机的非等时距滑坡位移预测模型不仅预测精度高，预测误差较小，且寻优参数设置简单，计算收敛迅速.

Prediction of non-equidistant landslide displacement time series based on grey wolf support vector machine

The Baishuihe Landslide in the Three Gorges Reservoir Area was taken as an example, and the non-equidistance and the complexity of the landslide displacement monitoring data were considered. A new non-equidistant displacement prediction model was proposed by combining the non-equidistant time series analysis, the grey wolf optimization (GWO) and the support vector regression (SVR). The natural cubic-spline interpolation method was employed to process the landslide displacement data. Then the trend component and the periodic component in landslide displacement were separated based on the time series approach. The cubic polynomial fitting based on robust least squares and the GWO-SVR coupling model were used to predict the trend displacement and the periodic displacement displacements respectively. The predicted value of landslide cumulative displacement was obtained through the time series model. Results show that the non-equidistant landslide displacement prediction model based on the grey wolf algorithm optimized support vector machine not only has high prediction accuracy and small prediction error, but also has simple calculation parameter settings and fast convergence.