基于MK-SVM和时序特征分析的月径流预报模型

Monthly runoff prediction model based on MK-SVM and time series feature analysis

针对传统径流预报方法预报因子不确定性和预报模型复杂性问题,基于月径流时序特征重要性分析选择预报因子,采用混合核函数支持向量机(MK-SVM)模型捕捉径流时序间的非线性关系,提出动态透镜成像反向学习和Lévy飞行等多策略融合的改进灰狼优化算法(IGWO),并构建了径流预报的IGWO-MK-SVM模型。黑河流域莺落峡水文站月径流预报结果表明:IGWO-MK-SVM模型月径流预报结果的纳什效率系数、均方根误差、Kling-Gupta效率系数分别为0.8942、16.9099 m^(3)/s和0.8639;与传统SVM模型相比,IGWO-MK-SVM模型在径流预报中的自适应性有所提升,相较于长短期记忆网络模型和季节性差分自回归移动平均模型,IGWO-MK-SVM模型能更好地预报月径流的真实变化过程。

To address the problem of uncertainty of prediction factors and model complexity of traditional runoff prediction methods,prediction factors were selected based on feature importance analysis of monthly runoff time series,and the nonlinear relationship between runoff time series was captured by the mixed kernel function-support vector machine(MK-SVM)model.An improved grey wolf optimizer(IGWO)that integrated multiple strategies,such as dynamic lens imaging reverse learning and Lévy flying strategies,was proposed to enhance the stability of the global parameter optimization of the MK-SVM model,and an IGWO-MK-SVM model for runoff prediction was constructed.The results of monthly runoff prediction at Yingluoxia Hydrological Station in the Heihe River Basin show that the Nash-Sutcliffe efficiency coefficient,root mean squared error,and Kling-Gupta efficiency coefficient of prediction results of the IGWO-MK-SVM model were 0.8942,16.9099 m^(3)/s,and 0.8639,respectively.Compared with the traditional SVM model,the IGWO-MK-SVM model has high adaptability in runoff prediction,and compared with the long short-term memory network model and the seasonal autoregressive integrated moving average model,the IGWO-MK-SVM model can better predict the real change process of monthly runoff.