大坝安全监测的兽棒最小二乘 支持向量机模型

针对目前在大坝监测模型中应用较多的支持向量机模型，以土坝沉降监测实例比较分析了监测数据中是否含有异常值的两种情况的最小二乘支持向量机监测模型的拟合精度与预测精度，发现异常值的影响不容忽视。通过改进支持向量机模型中的损失函数，建立了大坝安全监测的普棒最小二乘支持向量机模型(RLS一SVM )。实例分析表明:不论监测数据是否含有异常值RLS一SVM均可达到较好的拟合精度和预测效果，优于普通LS一SVM模型。

Robust least square support vector machine based model for dam safety monitoring

So far as the support vector machine based model frequently applied to the modeling of dam monitoring at present is concerned,a comparative analysis is made on the fitting precision and prediction accuracy of the least square support vector machine based monitoring model under the both the conditions with or without abnormal values in the monitoring data by taking the settlement monitoring on a earth dam as an actual study case,and then,it is found that the impact from the abnormal values cannot be ignored.Therefore,through the improvement of the loss function of the support vector machine based model,the robust least square support vector machine based model(RLS-SVM)for dam safety monitoring is established.The analysis on actual case shows that better fitting precision and prediction effect can be obtained from the model(RLS-SVM),no matter whether abnormal values are there in the monitoring data or not.Thus,the precision of RLS-SVM model is higher than that of LS-SVM model.