Abstract:

According to econometric time series model, the paper creatively proposed combing econometric model and fractal dimension based on Shanghai and Shenzhen stock market data system of great wisdom and statistical software EVIEWS during 2005-2009.It uses a high-dimensional chaotic characteristics of the stock index, using LP algorithm to determine the fractal dimension, using vector autoregressive VAR model to test unit root on the two stock markets in Shanghai and Shenzhen. According to AIC and SC information criterion, it determine the number of lags, predict the future trend of the stock market in dynamic and static state and obtain more reasonable results.

Key words:  composite index; fractal dimension; vector autoregressive VAR; forecast