The Research Of BP Neural Network In The Stock Price Forecast

Since the establishment of the stock market,stocks have become one of the financial management methods for most investors.The stock market is not only affected by macroeconomics,monetary policy,economic cycles,and the performance of listed companies.It is also a process of dynamic gaming between funds,technology,information,and talents among different interest groups based on the long and short sides.Is a nonlinear chaotic interaction process.With the gradual improvement of the market mechanism and the increase in the number of institutions participating in the stock market,China's stock market will form a large pattern of games among different institutions.As a result,many people participated in the analysis of the stock market and the prediction of the stock price.There are many types of stock forecasting methods.According to the previous perspective,there are three forecasting methods.These three methods will be introduced in detail in the third chapter of this article.The main method used in this article belongs to the technical analysis method.It builds mathematical models to predict time series data,and then makes recommendations to investors.This article first collates the domestic stock forecasting methods in recent years.Through comparison,it is found that domestic forecasting methods are mainly divided into two categories: the first is to apply statistical methods to forecast time series data;the second is to use neural Artificial intelligence models such as network models predict time series.The methods in statistics introduced in this article are some of the most basic analysis and forecasting methods in statistics,including regression analysis and time series analysis;these two types of analysis models make a simple prediction of stock data.The following describes how to apply the BP neural network time series training set for prediction.Because BP neural network has the characteristics of non-linear mapping,it can approximate the functional relationship with arbitrary precision,so it can fit well under the classified data sequence.This article is mainly to build a standard BP neural network model.By analyzing the convergence speed of the model,different parameter combinations are selected to determine whether the degree of fit is better.Record the number of hidden layers,the number of nodes,the training function,and the transfer function determined in the above process,and then apply genetic algorithms to optimize the initial parameters of the BP neural network and then predict the original time series data.Compared with all previous results,it is found that the GA-BP neural network model is better than the BP neural network model in predicting stock prices,that is,the GA-BP neural network model is more suitable for time series prediction.