An Artificial Intelligent Trading Support System For Taiwan Stock Market---Using Neural Network Based On Trading Analysis

• Main Authors: Duh, Shyh-Chorng, 杜世崇
• Other Authors: Ferng Wuu-Shyong
• Format: Others
• Language: zh-TW
• Published: 1998
• Online Access: http://ndltd.ncl.edu.tw/handle/37288528368482030614

碩士 === 國立臺灣大學 === 電機工程學系研究所 === 86 === Recently , because the improvement of both theory in neural network and functi on of personal computer , there are more and more examples using artificial in telligent in financial market. Using artificial intelligent not only can help us to deal with large information , but also can reduce the error caused by hu man emotion. In this paper, we hope that we can use artificial intelligent in Taiwan Stock Market from the aspect of market operation process. To use artifi cial intelligent in financial market, we have to concern about two major problems : (1)How to build the analysis model for the system ? and (2)What ki nd of artificial intelligent we need ? In the first problem , we build our ana lysis model based on characteristics of Taiwan Stock Market. In the second pro blem, we use two kinds of neural network : Adaptive Resonance Theory Network( ART-2) and Backpropagation Network (BPN) according to the analysis model we bu ild. We also make some improvements to accelerate the operation of entire syst em.In thispaper, we use the data of Taiwan Stock Market date from 1997.12.12 t o 1998.5.8 as the input of system , and try to forecast the stock price trend for ten stocks. In our simulation , we get better profit return than the raise of Taiwan Stock Market Index.