Stock Forecasting System Based On Bayesian Network Research And Design

The stock market securities and financial industry is an essential component in more and more China by all kinds of people's attention,the changes of stock price forecasting has become a hot topic,and has been a hot field of economics and mathematics research object.People need to collect,sort and analysis of very large amounts of information in the stock investment,including the macro field and micro field information,and the information of the market and non economic factors,the amount of information is very large and complex.In this paper,we use the principle of Bayesian network to establish the prediction model of stock volatility,and according to the actual demand of stock forecasting,we put forward a complete design scheme of stock forecasting reference system.First summarizes the influencing factors of stock volatility and the threshold factor summary,and use the average mutual information in information theory to quantify the influence factors of interrelated degree,whether a dependency exists between to determine the factors,as the prediction model of the initial construction of foundation.Secondly,to construct a Bayesian network structure forecast model of stock and using the expectation maximization algorithm study for the prediction model parameters using genetic algorithm,and to implement and optimize the algorithm,function demand analysis,stock forecasting system technical requirements,environmental requirements and put forward the overall goal of the system,which contains the module and process design,database design and system design system design.Finally,a series of experiments are carried out with real data,which proves that the prediction model can be constructed,studied and predicted effectively.