摘要
近年来,支持向量机(SVM)在建立大坝安全监控模型中得到了广泛应用,但其拟合精度和泛化能力取决于模型核参数σ和惩罚因子C的选取。以最小k-fold交叉验证误差为目标,用蚁群路径优化选择的节点值体现,并选择支持向量回归机中的核参数σ和惩罚因子C。以此建立了基于蚁群算法优化支持向量回归机(ACO-SVR)的混凝土坝变形监控模型。通过工程案例研究,结果表明:该方法在参数优化方面具有较快的寻优速率,用于混凝土坝变形监控的建模研究精度较高,具有广泛的适用性。
Recently, Support Vector Machine （SVM） has been widely used in the establishment of the dam safety monitoring model. However, because that its regression accuracy and generalization performance depend on a proper setting of relative parameters, a new trial applied in parameters selection called ACO-SVR is established based on the Ant Colony Optimization （ACO） Algorithm. The k-fold cross-validation error is used as the fitness function of ACO. The node values in the ant system are reflected by the kernel parameter cr and regularization pa- rameter C of SVR. Taking a concrete dam as an example, the fitness and the rationality in concrete dam deformation monitoring model is checked. Simulation results show that the optimal selection approach based on ACO has improved its optimizing speed. The research method of the concrete dam deformation monitoring model obtains a higher accuracy with a wide range of applicability.
出处
《中国农村水利水电》
北大核心
2017年第5期37-41,共5页
China Rural Water and Hydropower
基金
国家自然科学基金项目(51579083
51479054
41323001)
中央高校基本科研业务费专项资金项目(2015B25414)
关键词
混凝土坝
安全监控
回归型支持向量机
蚁群算法
concrete dam
safety monitoring
support vector machine for regression
ant colony algorithm
