挖掘量化關聯規則以預測台灣50成分股票漲跌之研究

CCUR > College of Business > Department of Information Management & Graduate Institute of Information Management > Thesis > Item 987654321/28447

Please use this identifier to cite or link to this item: https://irlib.pccu.edu.tw/handle/987654321/28447

Title:	挖掘量化關聯規則以預測台灣50成分股票漲跌之研究 Research on Mining Quantitative Association Rules to Predict Stock Price Changes for Firms Listed in Taiwan 50 Index
Authors:	朱光明 Gunawan, Dennis
Contributors:	資訊管理學系
Keywords:	Quantitative Association Rules Candlestick Pattern Stock Trend Forecasting Data Mining Taiwan 50 Index
Date:	2014-06
Issue Date:	2014-10-07 16:09:58 (UTC+8)
Abstract:	As stock market investing attracts many investors to get a profit, trend and stock price forecasting plays an important role in the stock market. Many professional traders use the technical analysis to analyze the stocks and predict the stock price or trend. Candlestick chart analysis is one of the technical analyses which are widely used for investment decision making. Nevertheless, the accuracy, when it is applied for companies listed in Taiwan 50 index as a stand-alone variable, is not as expected. In this research, the quantitative association rules which are used in decision making are generated using the trade volume, 5-day and 10-day moving average, price range, 1-day or 2-day candlestick pattern, and price change on the next day or the next fifth day. The data used in this research is stock prices of 38 companies which are listed in Taiwan 50 index from January 5th 1999 until February 27th 2014. The net profit which can be gained using one-day candlestick patterns with increase rules and decrease rules respectively is 122.075% and 45.6012%, whereas the net profit for two-day candlestick patterns with decrease rules is 84.2994%.
Appears in Collections:	[Department of Information Management & Graduate Institute of Information Management] Thesis

Files in This Item:

File	Description	Size	Format
fb141007160945.pdf		4963Kb	Adobe PDF	826	View/Open

View Licence

Loading...