應用類神經網路探討股市技術指標之有效管理

• Main Authors: Tzeng,Jia-Shiang, 曾家翔
• Other Authors: 李俊德
• Format: Others
• Language: zh-TW
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/12487208423752728324

碩士 === 大葉大學 === 資訊管理學系碩士班 === 96 === In this paper, TAIEX as a study, from the current Technical indicators to predict the stock market to test the accuracy . Research is used by the Deviation rate (BIAS), MACD, William indicators (WMS% R) to forecast accuracy by back-propagation neural network of training, analysis of a large number of historical data, to determine the future trend of the stock market, and also discuss the theory of efficient market hypothesis reliability. The results of this research show that a single technical indicators using neural network's overall accuracy rate can reach more than 70 percent , the appropriate combination of different indicators can reach more than 74 percent accuracy. If deduct part of the day unchanged could reach more than 83 percent accuracy rate. The results that the Technical indicators different set of parameters on the accuracy of stock price movements will be significant difference. This study provides a neural network to use the best of Deviation rate, MACD and William indicators are Technical indicators of the stock market forecasting system. Let stock market investors have a good reference tool to support investment decisions.