Applying Artificial Neural Networks in financial Market analysis by the Open Interest of TAIFEX

• Main Authors: Luo, Kuo-Hua, 駱國華
• Other Authors: Chen, An-Pin
• Format: Others
• Language: zh-TW
• Online Access: http://ndltd.ncl.edu.tw/handle/02015996638885911956

碩士 === 國立交通大學 === 管理學院碩士在職專班資訊管理組 === 97 === This study investigates the relationship between TAIEX futures and open interest of futures and options, using an integration of self-organizing map and back-propagation neural network-based trading model. The data of the trading volume and open interest of three legal persons , and the open interest of large futures traders are based on Taiwan Futures Exchange, which cover from July 2, 2007 to March 3, 2009. The review of literature summarizes that Taiwan stock market have the weak form efficiency. Moreover, many researches imply that institutional investors earn significant abnormal profit by exploiting their large investments, which could cause a change in the market trend. The purpose of this research is to analyze the trading behavior of institutional investors and large traders, and make further predictions of Taiwan stock index futures trend. The empirical results indicate that the open interest of three legal persons and large traders can be a factor of forecasting Taiwan stock index futures, besides we find that the integrate model of self-organizing map and back-propagation neural network is superior to the back-propagation neural network model, and determining the right threshold can lead to profitable growth in investments.