建構連動模式以預測泰國股市股票趨勢及評估投資報酬

CCUR > College of Business > Department of Information Management & Graduate Institute of Information Management > Thesis > Item 987654321/45043

Please use this identifier to cite or link to this item: https://irlib.pccu.edu.tw/handle/987654321/45043

Title:	建構連動模式以預測泰國股市股票趨勢及評估投資報酬 Constructing a Correlation Model to Predict Stocks Movement and Evaluate Investment Return of Thailand Stock Market
Authors:	林達財 RUNGROJKANRANAN, EKASIT
Contributors:	資訊管理學系
Keywords:	Correlation Analysis Data Filtering Stock Movement Forecasting Stock Co-movement Thailand Stock Market
Date:	2019
Issue Date:	2019-09-19 10:47:59 (UTC+8)
Abstract:	Investing in the stock market remains one of the most popular investment strategies from the past to the present. Many researchers are using data mining techniques such as neural networks and support vector machines to develop stock forecasting models. They tend to be very focused on predicting the price of a single stock based on their own data. This study proposes a stock correlation model to find the co-movement between two stocks by using the movement of stock A to forecast the movement of stock B which is delayed by 1-3 days. The correlation analysis generated stocks with highly correlated and negative pairs, then investment simulation model is used to generate buy or sell signal based on its time delay. The financial sector of the stock exchange of Thailand is measured in this experiment. The result shows that stocks with a correlation above 0.5 has the highest accuracy rate of gain at 68% and generates a 3.02% average profit per transaction. The stocks with a correlation between 0.4 to 0.5 show 56% of accuracy rate and 1.90% of average profit per transaction. Negatively correlated stocks show 67% of gaining accuracy rate and 1.15% of average profit per transaction.
Appears in Collections:	[Department of Information Management & Graduate Institute of Information Management] Thesis

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	142	View/Open

View Licence

Loading...