A Study of Looking for Low Risk Portfolios with Data Mining Technology in Taiwan Stock Market

• Main Authors: Sheng-chang Wu, 吳聲昌
• Other Authors: none
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/68390490902449963797

碩士 === 世新大學 === 資訊管理學研究所(含碩專班) === 94 === To invest stocks is a kind of general and important financial activities nowadays, and all of investor hope to trade with security, to raise the returns on investment and furthermore, to control the risk. Thus, in past, there were a lot of academic researches to study how to anticipate the stocks price fluctuation accurately to reduce the investment risk indeed and raise the rewards. In general speaking, there are two kinds of the research aspects in stocks price fluctuation anticipation. One is the basic analysis and anther is the technological analysis. In basic analysis, they think the stocks price fluctuates according to the company’s operation situation. So, they usually study from the company’s financial and try to find out the relationship between the financial statement of the personal share and stocks price. However, only studying the financial statement of the personal share is not enough, because the external overall economic environment also could be affected the stocks price. Thus, they should consider both of the external overall economic environment tendency and internal company’s operation situation, then, they just can anticipate more accurate for stocks price fluctuation. As for technological analysis, they are used much frequently by people in the stock market, but the variety and complexity of the technical indicator often makes the investors who lack the professional knowledge hesitate when they face the choice and lose the best opportunity to buy and sell their holding. So, it is also not a perfect way to be used in the stocks price fluctuation anticipation. The neural network is burgeoning information treatment technology and getting popular in recent years, especially using in financial data analysis has a good effect. So, this research is used by the pouring transmission neural network which is a kind of neural network methods to make a fast and effective sieving from numerous technical indicators, and combine with personal share chip analysis and positive & negative behavior in forward market, to predict the price of the opening quotation and closing of the next day of the personal share. It can help the investor to make the choice in the technical indicator and draft the investment tactics, so investor can reduce the risk of holding.