| As a member of the WTO, China will open industries of trade, finance, investment etc to other WTO members according to the bilateral WTO agreements. That makes the business circumstances of domestic enterprises more and more part of the regional economy or even the global economy . The close connection between one entity and the global economy - the open system will, no doubt, complicate the business circumstances of domestic enterprises and their behaviors .Since 1998, Economic and financial crises have been experienced in certain areas and countries,which are more or less related to economic opening ; as a result ,enterprises within such systems have suffered unpredictable losses. To meet the new business environment, consolidation of quantitative analysis and forecast is an indispensable and important component of routine and strategic management. It is obvious that the enhancement in these fields will contribute to enterprises' ability in analysis and prediction of opportunities and crises . Among data related to enterprises , one kind called time series, which reflects long term evolution of certain quantitative indexes of enterprise and tendency of macro economy. The former are annual profit rate, monthly sales growth etc , the later includes stock index, exchange rate ,interest rate etc . Study on such kind of data will benefit enterprises' operation analysis , investment , research and development ,international trade, prediction of macro economic states as well as identification of opportunities and crises.The paper concerns study on processing ,analysis and prediction of such kind data . Component stock index of Shenzhen stock market is taken as sample and the practical analysis is based on the theories of time series analysis and nonlinear complicated system. Interesting results gained in the paper :Brief review of development of Shenzhen stock market in the last 10 years with its component index shows the market has experienced course from the initial to related mature . The listed corporations in the stock market shall implement new strategy in funding or enterprise merger etc as the market develops .Based on sample of the index from April 3,1991 to May 31, 2001 ,ARIMA models have been built with TSP computer software guided by route of "from general to specific". The models built have better fit goodness and one point forward prediction is highly precise .but ultra-sample prediction by C++ program shows prediction precision reduces fast as the length of prediction grows ,long term prediction of the index is impossible . Spectral analysis of the data indicates long term factors dominate short term factorsin Shenzhen stock market , research in this aspect needs deepening .Results of nonlinear analysis shows : the maximum Lyapunov index of Shenzhen component stock is positive, which identifies chaos in Shenzhen stock market; correlated dimension of the system is between 1.7~2.2 , so system of Shenzhen stock market can be explained at least with 3 independent variables . Hurst index is bigger than 0.5,which proves system of Shenzhen stock market with property of fractal . From graph of Hurst index it can be seen the average period of trajectory in Shenzhen stock market is about 200 trading days, that indicates present stock index is related to indexes within 200 trading days backward. The nonlinear model built is suitable for the prediction in the short term, and ultra-sample prediction into 21 points forward has better precision, but long-term prediction is almost impossible because of the chaotic property.Comparison shows nonlinear model with theory of nonlinear complicated system is more precise than models of "classic time series analysis", that proves data with chaos shall be processed with theory of nonlinear complicated system and nonlinear models shall be established accordingly.Time series in enterprises shall be classified into two kinds, for purpose of data exploration. criterion for the work are "random" and "chaos ". Identification of data ty...
|
PDF Full Text Request