Yuyang Zhu, Linxian Zhi, and Weiduo Zhang
Stock price forecasting, artificial neural network, R-ELM, tappeddelay units
In this paper, a single-layer feedforward neural network (SLFN) model with tapped delay units at input layer is proposed to predict the daily closing price of the S&P 500 index. It is shown that by adding tapped delay units at the input layer, the dynamics in financial data can be captured effectively. In addition, training with the regularized extreme learning machine (R-ELM) method, not only the fast training speed is ensured but also the robustness of the neural model is improved. The performance of the proposed model is evaluated on the S&P 500 index raw dataset. The comparisons with backpropagation neural network (BP-NN), linear regression (LR), ARIMA and GARCH are conducted. The simulation results validate the efficiency and effectiveness of the proposed model.
Important Links:
Go Back
