Traditional statistical models work on the assumption of linearity and stationary of time series. Machine learning models such as Artificial Neural Network (ANN) and Support Vector Regression (SVR) suffer the problem of over-fitting and are sensitive to parameter selection, respectively. This paper propose two models that integrates Statistical Empirical Mode Decomposition (SEMD) and ANN and Ensample Empirical Mode Decomposition and ANN for improve the weakness of ANN. SEMD and EEMD are an adaptive technique that shifts the non-stationary and non-linear time series data till it becomes stationary. In first stage, the data is decomposed into a smaller set of Intrinsic Mode Functions (IMFs) and residuals using EEMD and SEMD. In the next stage, IMFs and residue are taken as the inputs for the ANN model. The methodology was compared with EEMD-ANN and SEMD-ANN models. The results suggest that the SEMD-ANN is better than EEMD-ANN.
展开▼
