The Mixed Model Through Multivariate Time Series Analysis And Lag Cointegration And Its Application In Prediction Of Stock Index

Stock index forecasting has always been a complex and difficult problem, on the one hand, because of the complex nature of the stock market,it is difficult to have accurate and comprehensive forecast of it, which is the charm of stock index prediction; on the other hand, at home and abroad, many scholars continues to use a variety of different methods and models to predict the stock index, with the deepening of the study, the predicted precision and prediction accuracy of those methods and models are greatly improved. However, the previous prediction models and methods have some problems that affect the accuracy of stock prediction and accuracy. Therefore, In this paper, the mixed alert model through multivariate time series analysis and lag cointegration is builded, and use it to predict the stock market, to some extent, which overcome the problem of considered incomplete of the time-series approach, and overcome the defect of time affect of the cointegration. In order to get higher accuracy and precision prediction.In theory, this article first introduces the basic concepts of time series analysis and models, discusses the basic steps of the establishment of time series analysis models, and then introduces the cointegration analysis; discuss the estimation and testing of cointegration. And leads to the concept of lag cointegration finally, the parameter estimation and testing of the lag cointegration are discussed.In empirical, on the basis of the theoretical analysis, the mixed alert model through multivariate time series analysis and lag cointegration is builded, The Shanghai Composite index as the object of study, select the broad money supply, savings deposits, in January of interbank lending rate, the RMB against the U.S. dollar exchange rate, industrial added value, total retail sales of social consumer goods, macroeconomic climate index, consumer price index, corporate goods price index and other indicators of macroeconomic variables as explanatory variables, using the monthly data from January 2000 to August 2010, first lag analysis to determine the optimal lag order of the explanatory variable and explained variables, and then timing analysis, onward multi-period of the time series which will participate in forecasting model to predict the explained variables; then the cointegration analysis, estimate the parameters of cointegration; finally the comprehensive analysis, estimates the forecast model.In the short term prediction results, the predicted model obtained good results, and for long-term forecast, the model parameters can be updated according to the real-time information to obtain better results.