Research On Stock Forecasting Techniques Based On Neural Network And Genetic Algorithm

Forecasting is an important link in scientific management and premise before policy-making and layout. It is necessary to forecast and analyze evolution trend of some systems. And stock market is an alarm, whose function is not only valued by the government but also concerned by investors.Artificial neural network is an embranchment of artificial intelligent, which was originated in 1940s and has been widely applied to many fields now. Neural network can reserve preening information and knowledge, which is the theoretical basis when used to forecast the future. As for time series forecast, neural network is more efficient and precise than mathematical models. Furthermore, with the great searching ability, GA (genetic algorithm) can solve many problems that other algorithms can't do. GA was combined with ANN, and improved the capability of the ANN greatly. The combination makes the Intelligent Computing develop rapidly. So, studying the two subjects is very significant.This paper starts with the basic theory of GA. The thesis presents GA--BP algorithm based on the high-nonlinear speciality of stock market, and also takes the shortcoming of basic BP algorithm that includes the slow convergence speed and local extremum into acount. The GA--BP algorithm has the character of the whole space search. The GA pays attention to the unknown-area search. At the same time, it also has high speed and relatively high precision. It will not get into local extremum. The BP algorithm searches the area that includes the whole space minimum. It can improve the speed and precision. Theoretical analysis and experiment results show that the method of stock prediction using neural network is feasible and efficient. It has favorable foreground. It is also proved that GA--BP algorithm can improve the speed and credibility.