Publication Type : Conference Paper
Publisher : Elsevier B.V.
Source : Procedia Computer Science, Elsevier B.V., Volume 143, p.947-953 (2018)
Keywords : Commerce, Convolutional Neural Networks (CNN), Deep learning, Electronic trading, Extreme learning machine, Financial markets, Financial time series, Forecasting, Long short-term memory, LSTM, Mean square error, Root mean squared errors, Stock price forecasting, Stock price forecasts, Stock price movements, Time series
Campus : Coimbatore
School : School of Engineering
Department : Electronics and Communication
Verified : Yes
Year : 2018
Abstract : Stock price time series are extremely nonlinear in nature and hence, accurate stock price forecasting has been a challenge. Accurate prediction of stock prices and the direction of stock price movement is also essential for a stock trader/investor in order to trade profitably. A deep learning approach to stock price forecasting is presented in this study. A total of fourteen different deep learning models based on Long-Short Term Memory (LSTM), Gated Recurring Unit (GRU), Convolutional Neural Networks (CNN) and Extreme Learning Machines (ELM) are designed and empirically evaluated on all stocks in the S&P BSE-BANKEX index for their ability to generate one-step ahead and four-step ahead forecasts. Performance of the proposed systems is evaluated in terms of the Root Mean Squared Error (RMSE), Directional Accuracy (DA) and the Median Absolute Percentage Error (MdAPE). Results indicate that deep learning models proposed in this study are capable of generating highly accurate stock price forecasts. © 2018 The Authors. Published by Elsevier B.V.
Cite this Research Publication : J. A. Balaji, Ram, D. S. Harish, and Dr. Binoy B. Nair, “Applicability of deep learning models for stock price forecasting an empirical study on bankex data”, in Procedia Computer Science, 2018, vol. 143, pp. 947-953.