Predicting US Stock Prices Using Long Short-Term Memory (LSTM)

Dayang Afiqah Liyana, Abang Ehsan (2023) Predicting US Stock Prices Using Long Short-Term Memory (LSTM). [Final Year Project Report] (Unpublished)

	PDF Dayang Afiqah Liyana (24pgs).pdf Download (1MB)
	PDF (Please get the password by email to repository@unimas.my , or call ext: 3914 / 3942 / 3933) Dayang Afiqah Liyana (fulltext).pdf Restricted to Registered users only Download (11MB)

Abstract

A stock, commonly referred to as equity, represents an investment that signifies partial ownership in a company. Investors are concerned with two crucial aspects: the current price of their existing or potential investment and its projected selling price in the future. Predicting stock prices has always been of great interest to investors; however, it has proven to be a challenging task for researchers and analysts. The stock market is highly unpredictable, with numerous complex financial indicators. Consequently, financial analysts, researchers, and data scientists are continuously exploring analytical tools to uncover stock market patterns. In this study, historical stock price data is leveraged to predict the stock prices of selected US companies using a machine learning approach known as the Long Short-Term Memory (LSTM) Model, which is a specialized form of Recurrent Neural Network (RNN). The dataset comprises five years of AAPL and MSFT data obtained from Yahoo Finance, with consideration given to six relevant attributes. The LSTM model is employed to generate accurate and reliable predictions. The LSTM model holds several advantages in the realm of stock price prediction as it utilises historical stock price data to discern patterns and trends, enabling the forecasting of future price movements. This study focuses on employing the LSTM model to shed light on the potential for achieving precise stock price forecasts using machine learning techniques. Additionally, the Root Mean Square Error (RMSE) is employed as a supplementary performance measure alongside the LSTM model for stock price prediction

Actions (For repository members only: login required)

View Item