Stock Price Trend Forecasting using Long Short Term Memory Recurrent Neural Networks

Abstract

The prediction of future stock price trend using current and historical stock market data is a research problem for traders and researchers. Recently deep learning methods shown promising performance to extract meaningful information from the given large data. In this paper, we proposed a system to predict the next trading session close price trend from historical stock trading data using long short term memory (LSTM) method. This is a classification problem next trading session close price trend can be uptrend, downtrend, or sideways trend. We built an automated trading system using the results of our classifier. We experimented with the proposed trading system on the American index stocks. Our experimental results show that the proposed method outperforms the buy-and-hold and decision tree-based method.