Prediction Of Particle Swarm Optimization Neural Networks In Various Stock Markets

With the development of China’s stock market, stock market investing activities gradually become more frequent, the stock market has gradually become the most active market in the securities market, stock has become the most popular product of investors. Therefore, the prediction of the stock price has become a hot research area. Effective prediction analysis methods can help investors to develop investment strategies, increase revenue while reducing risk. The stock market is a very complex system, but its internal law has certain trends, and it is affected by many of the economic, political and other factors, but still the law of motion of the stock market system is still difficult to grasp. Many scholars has research on the stock market, and many model are built. Traditional research methods are mainly based on the mathematical model of statistical theory, like time series model has many applications in this field, but the prediction accuracy can not reach people’s requirements,in fact that people has gradually discovered the stock market system is a complex nonlinear system in the study, the traditional linear model can not be a good approximation of its internal rules, many scholars began to study the chaotic nature of the stock, and with the nonlinear algorithm’s development, many scholars began to use nonlinear algorithm neural networks, like genetic algorithms to predict the stock market,many stock prediction model based on the nonlinear algorithm is established.The article mix Particle Swarm algorithm and bp neural network, ues Particle Swarm algorithm to optimize the training method of neural network’s weight and threshold, discuss the selection and optimization of various parameters, establish the particle swarm optimization bp neural network model and use in the research for stock prediction. By three empirical analysis of the three representative index of the market, and comparision of prediction result of the particle swarm optimization neural network and the traditional bp neural network prediction show that the particle swarm algorithm can effectively strengthen the neural network predict capability, reduce the prediction error and improve the speed of training; prediction results of the three major markets is nice so it shows the stock market’s predictability; compare to the other two markets,U.S. stock market havs a more predictability.