Prediction And Analysis Of Stock Market Based On Improved GARCH Family Modeland Neural Network

Foreign stock market has a history of several hundred years since it first appeared.But the development of domestic stock market is only in a few decades.With the improvement of domestic economic level,the domestic stock market is also on the rise.But in recent years,with the change of the international situation,the imperfection of the domestic stock market mechanism,the incomplete understanding of the national stock market investment and other reasons,the domestic stock market often fluctuates greatly.This not only directly affects the economic development of our country,but also strikes people's confidence in the stock market.People gradually realize that the stock market is a market full of risks.People are eager to control risks while pursuing high returns,which requires forecasting the stock market.Scholars at home and abroad have put forward a lot of stock marketforecasting methods,which is of great significance for the research and improvement of these forecasting methods.The first three chapters mainly introduce the research background of stock market and the development history of domestic and foreign stock market.It also introduces the basic knowledge of the stock market,including the analysis of the basic and technical aspects of the stock market,some commonly used terms and indicators in the stock market and the commonly used methods of predicting the stock market.Chapter four and chapter five focus on the improved GARCH family model theory and BP neural network theory.In the sixth chapter of this paper,we will use these two methods to analyze the two typical indexes of Shanghai Stock Index(sh000001)and Shenzhen Stock Index(sz399001)in the domestic stock market.First of all,in the traditional GARCH family model,this paper adds four factors: volume,price range,transaction amount and P / E ratio.By comparing the fitting results of model coefficients and the errors between the final and actual volatility values,the best prediction model for the two indexes is selected,and the predicted volatility image is drawn.Secondly,for the same prediction range of Shanghai index and Shenzhen index,this paper establishes a three-layer BP neural network with 50 nodes in the hidden layer.According to the experimental results of different iterative algorithms,the best algorithm is selected to predict the closing price of the stock market,and the trend images of Shanghai index and Shenzhen index are predicted.Finally,the two prediction methods are compared,and their advantages and disadvantages,as well as theneed for improvement are analyzed.