Comparison of BP Algorithm and Asymmetrical ARCH Model for Fluctuation Prediction of Shanghai Stock Market
In this paper, BP algorithm technology and asymmetrical ARCH model are utilized to predict the fluctuation of Shanghai stock market, adopting the closing price of each weekend of composite index of Shanghai stock market, amounted to 675 weeks from December 19, 1990 to April 30, 2004. Then 3 kinds of commonly used counting amount of prediction error are used to examine the prediction result of the data outside the sample. The result of study indicates that BP algorithm is superior to the asymmetrical ARCH prediction model; among the asymmetrical ARCH models, EGARCH (2, 5) model is superior to TARCH (1, 4).
Shaobo LIU Sulin PANG
Institute of Finance, Jinan University Guangzhou, China 510632 Department of Mathematics, Jinan University Guangzhou, China 510632
国际会议
青岛
英文
2006-07-21（万方平台首次上网日期，不代表论文的发表时间）