支持向量机在变形监测数据处理中的应用

Application of SVM to Deformation Monitoring Data Processing

支持向量机是基于统计学习理论的结构风险最小化原则的数据挖掘新方法,它将最大分界面分类器思想和基于核的方法结合在一起,具有很强的泛化能力,能保证所得解是全局最优解。文中简要介绍了支持向量机的基本原理及其在变形监测数据处理中的应用,论述了如何利用支持向量机进行建模和预报。通过对某大坝变形监测的连续观测数据的计算分析。

The support vector machine（SVM） is the latest data mining method based on the statistical learning theory and the structural empirical risk minimization principle. It integrates the thoughts of optimal separating hyperplane with the kernel function based methods has good generalization ability, and generates unique and globally optimal solutions. The issues on how to recognize the model, build a model and predict the utilization of monitoring serial are discussed. A case study is given with the continuous analysis on periodical settlement data of a dam deformation monitoring point. The result shows that this method is feasible and effective in the process of engineering deformation monitoring.