论文部分内容阅读
支持向量机(SVM)的核函数类型和超参数对边坡位移时序预测的精度有重要影响。鉴于局部核函数学习能力强、泛化性能弱,而全局核函数泛化性能强、学习能力弱的矛盾,通过综合两类核函数各自优点构造了基于全局多项式核和高斯核的混合核函数,并引入粒子群算法(PSO)对最小二乘支持向量机(LSSVM)超参数进行全局寻优,提出了边坡位移时序预测的混合核函数PSO-LSSVM模型。将模型应用于锦屏一级水电站左岸岩石高边坡变形预测分析,并与传统核函数支持向量机预测结果进行对比分析。结果表明,该模型较传统方法在预测精度上有了明显提高,预测结果科学可靠,在边坡位移时序预测中具有良好的实际应用价值。
The kernel function types and hyperparameters of Support Vector Machine (SVM) have an important influence on the accuracy of slope displacement prediction. In view of the strong learning ability of local kernel function, weak generalization performance, strong generalization performance of global kernel function and weak learning ability, a hybrid kernel function based on global polynomial kernel and Gaussian kernel is constructed by synthesizing the advantages of both kernel functions. Particle swarm optimization (PSO) is used to globally search the LSSVM hyperoptimal parameters. A hybrid PSO-LSSVM model is proposed for predicting slope displacement time series. The model is applied to the prediction and analysis of rocky high slope deformation on the left bank of Jinping I Hydropower Station, and compared with the prediction results of traditional kernel function support vector machine. The results show that the proposed model is more accurate than the traditional method in predicting accuracy and the prediction result is scientific and reliable. It has a good practical application value in predicting slope displacement time series.