EEMD与RBF神经网络的太阳黑子月均值预测
发布时间：2019-03-07 11:12
【摘要】：太阳黑子月均值是典型的混沌时间序列,具有较强的非线性和非平稳特征,能够反映太阳活动的真实水平。采用一种应用集合经验模态分解(Ensemble Empirical Mode Decomposition,EEMD)与径向基函数(Radial Basis Function,RBF)神经网络组合的预测模型。通过EEMD将原始时间序列分解为若干个不同时间尺度的本征模态函数(Intrinsic Mode Function,IMF)分量,并对这些分量进行建模预测,再将各分量的预测值重构得到原始时间序列的预测值,这样不仅降低了算法的复杂性,而且有利于提高模态分量包含信息的物理意义。仿真结果表明,与经验模态分解(Empirical Mode Decomposition,EMD)结合RBF神经网络的模型相比,该模型具有较高的预测精度。
[Abstract]:The monthly mean of sunspot is a typical chaotic time series with strong nonlinear and non-stationary characteristics and can reflect the true level of solar activity. A prediction model based on the combination of set empirical mode decomposition (Ensemble Empirical Mode Decomposition,EEMD) and radial basis function (Radial Basis Function,RBF) neural networks is proposed. The original time series is decomposed into several intrinsic modal function (Intrinsic Mode Function,IMF) components with different time scales by EEMD, and these components are modeled and predicted, then the predicted values of each component are reconstructed to get the predicted values of the original time series. This not only reduces the complexity of the algorithm, but also improves the physical meaning of modal component inclusion information. The simulation results show that compared with the empirical mode decomposition (Empirical Mode Decomposition,EMD) model combined with the RBF neural network model, the proposed model has a higher prediction accuracy than that of the empirical mode decomposition (EMD) model.
【作者单位】： 西安邮电大学电子工程学院;
【基金】：陕西省自然科学基金(No.2014JM8331)
【分类号】：P182.41;TP183
，
本文编号：2436064
[Abstract]:The monthly mean of sunspot is a typical chaotic time series with strong nonlinear and non-stationary characteristics and can reflect the true level of solar activity. A prediction model based on the combination of set empirical mode decomposition (Ensemble Empirical Mode Decomposition,EEMD) and radial basis function (Radial Basis Function,RBF) neural networks is proposed. The original time series is decomposed into several intrinsic modal function (Intrinsic Mode Function,IMF) components with different time scales by EEMD, and these components are modeled and predicted, then the predicted values of each component are reconstructed to get the predicted values of the original time series. This not only reduces the complexity of the algorithm, but also improves the physical meaning of modal component inclusion information. The simulation results show that compared with the empirical mode decomposition (Empirical Mode Decomposition,EMD) model combined with the RBF neural network model, the proposed model has a higher prediction accuracy than that of the empirical mode decomposition (EMD) model.
【作者单位】： 西安邮电大学电子工程学院;
【基金】：陕西省自然科学基金(No.2014JM8331)
【分类号】：P182.41;TP183
，
本文编号：2436064
本文链接：https://www.wllwen.com/kejilunwen/zidonghuakongzhilunwen/2436064.html
