The Fractal Analysis And Bp Neural Network’ Forecasn G Of China’s Main Stock Index

As is known, the stock market is extremely important in the national economy. what kind of rules does China’s stock market follow? The linear or nonlinear, the random model or the gaussian model or a fractal structure? Only thoroughly understand these problems can we analysis the market, forecast the market better, grasp the laws of the market and make better use of the capital market of our country.China’s stock market contains more than2000listed companies, as well as a variety of plates, index, etc.. Generally, people choose the Shanghai composite index and the shenzhen component index as representative to analysis China’s stock market because of the shanghai composite index represents the shanghai stock market, the shenzhen component index represents the shenzhen stock market. The article will compare the csi300index with the shanghai composite index and the shenzhen component index respectively. We found out that the shape of the csi300index is similar to the Shanghai composite index and the shenzhen component index. So we can chose the csi300index to analyze China’s stock market.This article gives a research on yield rate of the csi300index through the analysis of R/S. The price return is considered in day, week and month period. The results show that the Hurst index is more than0.6and draws the conclusion that the yield rate is not a normal distribution but a fractal structure. The chapter give the prediction interval of fourth chapter of the paper.The artical choose the BP neural network in predicting the csi300index and selecte EMA line to improve the original model. The results show that These two models can outperform the average yield of the market and the Improved model is more accurate. The maximum deviation error is4.69%of the positive numbers and the maximum deviation error is-3.73%of the negative numbers of the411data. According to the predicted results we find that updating time series of training samples in time can improve prediction accuracy.We can draw three important conclusions through this paper. Frist, the prediction model of the csi300index based on BP neural network is feasible and the model can outperform the market. Second, it is important to choose suitable prediction intervals. Third, China’s stock market is a fractal structure and its effectiveness is not evident. The market’ event-driven is evident. Predicting the China’s stock market is very important when quantifing the influencing news. So predicting the stock market successfully still has a long way to go.