Prediction of Short-term Stock Prices Based on EMD-LSTM-CSI Neural Network Method

In order to improve the accuracy and efficiency of short-term stock price trend prediction, a new prediction model based on Empirical Mode Decomposition, Long Short-Term Memory neural network and Cubic Spline Interpolation (EMD-LSTM-CSI) was proposed. Firstly, the stock price data series are decomposed into several Intrinsic Mode Functions (IMF) and a Residual component (RES) with different time scales. Then, in order to avoid over fitting, components are divided into two categories according to their gradient order of magnitude, and the larger one is recorded as the first category. The LSTM model is built for each component in the first category, and the CSI model is built for each component in the other category. Finally, the final prediction results are obtained by summing up the prediction series of each component. Compared with LSTM model, EMD-LSTM model and SVM model, the EMD-LSTM-CSI model can achieve better accuracy, which provides a more credible reference for investors.