Volume 12, Issue 7, July 2020



A Hybrid Arima-Lstm Model for Stock Price Prediction

A Hybrid Arima-Lstm Model for Stock Price Prediction

Pages: 48-51 (4) | [Full Text] PDF (529 KB)
UFI Abdulrahman, N Ussiph, B Hayfron-Acquah
School of Technology, Christ Apostolic University College, Kwaadaso, Ashanti, Ghana
Department of Computer Science, Kwame Nkrumah University of Science and Technology, Kumasi, Ashanti Ghana


Abstract -
The stock market offers investors the opportunity to trade in shares and equities. Making profit in the stock market depends on the ability to accurately predict the future stock prices. This is leading to the development of stock price prediction models using various methods such as ARIMA, LSTM, and Neural Network. Even though these models have proven to predict future stock prices, recent time is witnessing the development of better models using hybrid approaches. This study therefore propose a hybrid model by combining ARIMA and LSTM based on data decomposition with low-pass filter of the discrete Fourier Transform. Experiment carried out with data from the Ghana stock exchange reveals that, the hybrid model performs better than the individual ARIMA and LSTM models. This is confirmed with their R.M.S.E values.
Index Terms - Stock Market, stock prediction, Hybrid stock predictions, ARIMA, LSTM

Citation - UFI Abdulrahman, N Ussiph, B Hayfron-Acquah. "A Hybrid Arima-Lstm Model for Stock Price Prediction." International Journal of Computer Engineering and Information Technology 12, no. 7 (2020): 48-51.