Attention-Ordering LSTM for Financial Time Series Forecasting

Accurate forecasting in the financial field has been widely discussed. With the development of deep neural networks, it is also used in financial forecasting. In order to accurately predict financial time series, this paper proposes an attention-ordering long short-term memory model (AO-LSTM) combined with the empirical mode decomposition. As a popular deep recurrent network, long short-term memory can efficiently predict the financial time series. The preprocessing using empirical mode decomposition can improve the prediction accuracy. Experimental results show that the proposed prediction model obtains the best prediction performance.