Study on the Improvement of the Application of Complete Ensemble Empirical Mode Decomposition with Adaptive Noise in Hydrology Based on RBFNN Data Extension Technology

The complex nonlinear and non-stationary features exhibited in hydrologic sequences make hydrological analysis forecasting difficult. Currently, some hydrologists employ the complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) method, a new time-frequency method based on (EMD) algorithm, to decompose raw data order obtain relatively stationary components for further study. However, endpoint effect CEEMDAN is often neglected, which can lead errors that reduce accuracy of research results. In this study, we processed an original runoff sequence using radial basis function neural network (RBFNN) technique extension before utilizing decomposition. Then, compared results sequence, RBFNN standard investigate influence method. indicated effectively reduced error medium low frequency caused by effect. At both ends components, more accurately reflected true fluctuation characteristics variation trends. These advances are great significance subsequent study hydrology. Therefore, combined appropriate series, precisely determine multi-time scale characteristics, provide credible time series forecasting.