An improved LSTM technique using three-point moving gradient for stock price forecasting

Wong, Pee Shoong and Asmai, Siti Azirah and Zulkarnain, Nur Zareen and Tay, Choo Chuan (2022) An improved LSTM technique using three-point moving gradient for stock price forecasting. INTERNATIONAL JOURNAL OF COMPUTER INFORMATION SYSTEMS AND INDUSTRIAL MANAGEMENT APPLICATIONS, 14. pp. 338-346. ISSN 2150-7988

Text 2.3.1.2.2. AN IMPROVED LSTM TECHNIQUE USING THREE-POINT MOVING GRADIENT FOR STOCK PRICE FORECASTING.PDF Download (796kB)

Abstract

Everybody desires to see the future. The understanding of the time ahead can create enormous prospects, clout and fortune. For soothsayers who can predict the future financial developments, they are be able to command immense fortune. Therefore, predicting the financial stock market is always a great attraction with financiers and opportunists. Massive finance and equipment have been devoted into exploring the stock prices movements, hoping of realizing the best method to hit paydirt in this endervour. Analysts utilize numerous methodologies as well as employ a variety of hypotheses to solve this artful task, but not any has accomplished with a satisfactory margin of errors. The need to make informed choices causes correct and accurate information to be a desired and highly valued commodity. One encouraging Artificial Intelligence technique is the Long Short-Term Memory (LSTM) technique. Although LSTM is excellent in recognizing patterns, it places no considerable importance to the intense relationship that connects the past and its ensuing values within a time series. It does not take into account the association between the past and the later values. This research offers an optional LSTM method which contemplates the past, current and subsequent prices by predicting the Three-Point Moving Gradient of the stock market prices. The precise forecast price can be calculated with adequate accuracy by employing reversed Linear Regression (LR) and its mean price. The results is just as precise as the conventional LSTM method.

Item Type:	Article
Uncontrolled Keywords:	Time series forecasting, Long short-term memory (LSTM), Linear regression, Deep learning, Three point moving gradient
Divisions:	Faculty of Information and Communication Technology
Depositing User:	Sabariah Ismail
Date Deposited:	28 Mar 2023 13:55
Last Modified:	28 Mar 2023 13:55
URI:	http://eprints.utem.edu.my/id/eprint/26447
Statistic Details:	View Download Statistic

Actions (login required)

View Item