S.-I. Wu and H. Zheng (USA)
Artificial Intelligence Applications, Data Mining, Neural Networks, Time Series Analysis
We have analyzed the daily closing prices of stock indexes (S&P500, Dow Jones, and NASDAQ) to determine appropriate ARIMA models using time series analysis techniques. We have also developed recurrent neural networks to forecast the daily closing prices of these stock indexes. Experiments showed that the price is predictable and much better than the random guess. According to a simple short-term investment strategy, a good annual profit rate can be obtained. Different activation functions were tested in order to dynamically choose the best neural network topology. We explored more than 600 network structures for each neural network. Based on the ARIMA models developed, their autoregressive parameters helped us determine the number of recurrent neurons in recurrent neural networks.
Important Links:
Go Back
