大坝变形监测数据由于受水位、温度、时效等因素的影响,是一个非线性、非平稳的时间序列。考虑到支持向量机模型(SVM)对小样本、非线性问题有很好的预测效果,采用SVM模型刻画样本数据的趋势性信息;同时运用径向基函数神经网络(RBFNN)对残差序列进行分析和预测,并将时间序列的趋势项和误差项线性叠加,组成SVM-RBFNN组合预测模型。实例应用表明,SVM-RBFNN组合模型精度高于SVM模型,具有一定的实用性。 Dam deformation monitoring data due to water level, temperature, aging and other factors, is a non-linear, non-stationary time series. Considering that the SVM has a good prediction effect on small samples and nonlinear problems, the SVM model is used to characterize the trend of the sample data. At the same time, RBFNN is applied to the residual sequences Analysis and forecast, and linearly add the trend items and error items of the time series to form a combined forecasting model of SVM-RBFNN. The application of the example shows that the SVM-RBFNN combined model has higher accuracy than the SVM model, which has some practicality.