Stock prices time series forecasting by deep learning using three-point moving gradient

Wong, Pee Shoong and Asmai, Siti Azirah and Tay, Choo Chuan (2020) Stock prices time series forecasting by deep learning using three-point moving gradient. International Journal of Advanced Trends in Computer Science and Engineering, 9 (4). pp. 6622-6630. ISSN 2278-3091

[img] Text
00472050920241028571096.pdf

Download (1MB)

Abstract

Everyone wants to know the future. For the knowledge of the future bring immeasurable opportunities, power and wealth. Anyone who can foresee what the future economic trends are, can generate enormous wealth. Thus, forecasting the stock market has always been a fascination among investors and speculators. Billions of dollars and tons of equipment have been poured into researching the movement of stock prices, with the hope of discovering the ultimate formula to strike gold in this venture. Forecasters employ various methods and apply various theorems to crack this elusive undertaking, but none has really succeeded accurately producing forecasts with acceptable margins of errors. One promising AI method is the Long Short-Term Memory (LSTM) method. Though LSTM is good at pattern recognition, it does not give much consideration to the strong affinity that binds the preceding and the succeeding values in a time series. It does not consider the relationship between the preceding and the succeeding values. This study proposes an alternative Artificial Intelligence-infused Long Short-Term Memory method (LSTM) which considers the preceding, present and succeeding values by forecasting the Three-point Gradient of the stock prices. By using reversed Linear Regression and the mean price, the precise forecast price can be reversed calculated with acceptable accuracies. The results may be just as accurate as normal LSTM method or even more if correctly tweaked and may require less process time.

Item Type: Article
Uncontrolled Keywords: Gradient, Long short term memory, Linear regression, Stock prices forecasting
Divisions: Faculty of Information and Communication Technology
Depositing User: Sabariah Ismail
Date Deposited: 09 Oct 2024 16:28
Last Modified: 09 Oct 2024 16:28
URI: http://eprints.utem.edu.my/id/eprint/27831
Statistic Details: View Download Statistic

Actions (login required)

View Item View Item