Portfolio Management Using Artificial Neural Networks

• Main Authors: Jia-Yang Wu, 吳佳陽
• Other Authors: Chin-Sheng Huang
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/39910910897298327742

碩士 === 國立雲林科技大學 === 財務金融系碩士班 === 94 === The purpose of a portfolio makes assets to allocate for several superior underlying investment, which makes the largest expected return on an investment. The concept of an “efficient portfolio” has been developed by the portfolio theory, and focuses on portfolio selection. This study used artificial neural network (ANN) to analyze and forecast the stock price and its future trend, which construct a fitting portfolio to help the decision-making of investment. Back-Propagation Network (BPN) was used to perform empirical analysis of the thesis and the results show that the order of average returns for between 2000 and 2004 was first the market portfolio, followed by the ANN portfolio, and last the Markowitz portfolio. The order of average risk between 2000 and 2004 was first the ANN portfolio, followed by the Markowitz portfolio, and last the market portfolio. When the number of corporations in the portfolio was increased, the risk of the ANN portfolio declined, and the portfolio theory’s risk diversification principle was attained. This study shows that the ANN portfolio is superior to the Markowitz portfolio.