Comparison of Stock Market Prediction Using Deep Learning Algorithms

Abstract

AbstractThe stock market is the act of buying and selling the share of the companies and yield more profits. In order to provide the abnormal returns for the company by share market, the prediction of the stock market is necessary. Using deep learning algorithms, this analytics process predicts the daily return direction of the SPDR S&P 500 ETF (ticker symbol: SPY). The deep learning algorithms such as simple recurrent neural network and long short-term memory algorithm are applied to predict the daily direction of future price of S&P 500 based on the historical price. The performance of different procedures such as simple RNN and LSTM is compared. LSTM algorithm is found to be an accurate algorithm when compared to the other algorithms. The simple RNN and LSTM algorithm is implemented on another dataset such as Bombay Stock Exchange (BSE), and the performance of both algorithms is compared and simulated; the result of LSTM is a better algorithm, to predict a stock market daily return.KeywordsDaily stock return forecastingDeep learning algorithmsRNN—recurrent neural network and LSTM—long short-term memorySPDR S&P 500 ETF