Financial time series forecasting model based on CEEMDAN-LSTM

Currently, one of the most important problems in predicting non-stationary and nonlinear financial series is that the existing models are not effective enough. A new model based on complete ensemble empirical mode decomposition with adaptive noise(CEMDAN) and long short-term memory(LSTM) network is proposed to forecast the financial time series. In practice, future information cannot be used for prediction, so a sliding time window is introduced to decompose the original sequence into a cluster of equal length sequences. After CEEMDAN decomposition, wavelet threshold denoising and reconstruction, the denoised signal was obtained. Using the denoised signal instead of the original signal as the input of the LSTM network, the more accurate final prediction result will be obtained. Take the Ping An Bank’s share price data as an example, a variety of models are used to predict its price. The result shows that that the proposed CEEMDAN-LSTM model has better prediction effects than the standard LSTM model and the common prediction methods combined with empirical mode decomposition.