The Modeling And Application In Stock Market Forecast Based On Neural Network With Particle Swarm Optimization

For increasing the income and reducing the risk, investment activities in the stock market need an effective prediction and analysis method of stock market. Stock market is a highly complicated nonlinear dynamic system, its variation has its own regulation, but also is influenced by many other factors, such as politics, economy and psychology. While traditional quantitative prediction techniques based on statistics meet difficulties in stock market analysis, neural network enjoys the virtue of self-organization and self-adaptation, and can learn the economical knowledge from historical data. So it is suitable to solve problems in traditional methods of stock market prediction, and the research of stock market forecast based on neural network has certain theoretical significance and practical value.However, because of the limitations of the traditional neural network training method, it is difficult to improve the accuracy of neural networks prediction method. In order to improve the efficiency and accuracy of the stock market prediction, this paper proposed a hybrid algorithm combined with Particle Swarm Optimization (PSO) algorithm and Back Propagation (BP) algorithm after analyzed the problem of neural network research and stock market prediction at present. Particle Swarm Optimization algorithm is a new theory based on swarm intelligence. The algorithm can provide efficient solutions for complicated optimization problems by the intelligence generated from complex activities, such as cooperation and competition among individuals in biologic colony.This paper gave the general steps of stock market prediction using BP neural network, and set up a stock market prediction model based on BP neural network optimized by Particle Swarm Optimization algorithm. In order to solve the problem that China's stock market being influenced strongly by the macro-policy factors, the paper input environment variables in the model which reflect the influences of environmental factors. Using the model set up, the paper analyzed Shanghai Composite Index of China, and compared the outcome of forecast with it of traditional BP neural network.Empirical results indicated that the Particle Swarm Optimization algorithm improved effectively the tendency of BP network falling into partial optimum, increased greatly the convergence speed and prediction accuracy of BP neural network, and ameliorated the performance of algorithm and the results of prediction at a certain extent. The introduction of environment variables into the prediction model, improved the real-time quality of the network and enhanced the generalization ability of the network, and provided an effective new method for the stock market prediction.