A Study On Forecasting Of Stock Index Based On Wavelet Multi-resolution Analysis And GARCH Model

With the economic globalization speeding up, financial markets have become increasingly important in the economy of a country. As an important part of the financial market,the stock market attracted more and more attention of the country. Chinese stock market is now in the development stage.Compared with the foreign mature stock market,the fluctuation of China’s stock market is more intense, and our stock market is more easily affected by external environment.In this case the stock market risk will increase, and then affect the operation of the national economy, exacerbated investors investment risk. So we need to use scientific methods to analyze stock market volatility forecasting.With expanding and deepening research on stock market volatility forecasting,the researchers has developed a variety of different prediction methods according to different needs,such as fuzzy prediction method,wavelet analysis forecasting methods,ARCH model forecasting,GARCH model prediction,etc.In this paper, an improved method by combing the wavelet multi-resolution analysis and GARCH model is provided.Then we got wavelet analysis-GARCH model prediction method.We get a good prediction effect.This paper is divided into five sections.First introduced the stock price volatility trend forecast’s importance and significance to the country’s financial regulation and personal investment.Then analyzes the characteristics of the fluctuation of stock index.The second part introduces the theory of wavelet multi-resolution analysis and wavelet de-noising theory. The third part is the introduction of the ARCH models. The fourth part is the application of wavelet analysis-GARCH model forecasting method.Select the CSI 300 index closing price time series as the research sample for empirical research. The last part is the relevant conclusions and discussion.Research process and conclusions show : Compared with the Dow Jones Industrial Average index,the volatility of CSI 300 is stronger. In the process of model selection can be seen,the CSI 300 Index has a certain "leverage effect".Asymmetric information exists impact.The prediction process indicates that using the wavelet multi-resolution analysis-GJR-GARCH model,the prediction accuracy is improved obviously.This indicates that the wavelet multi-resolution analysis de-noising on the improvement of China’s CSI 300 Index is effective.