Stock Price Forecasting Based On BP Neural Network Optimized By Genetic Algorithm

With the development of social economy, more and more people are attention toinvestment and financial management. To invest stock market is their number onechoice. However, stock market that can be influenced by many factors, such aseconomic factor, policy factor and investor’s psychology etc is a kind of highlycomplex nonlinear dynamic system. Due to the uncertainness of the stock market,stock investment is a model of investment with high earning and high risk. Onceinvestors forecast are not accurate, it will cause capital losses. Effective analyticalmethods to assist in the decision are needed for investors, so that risks are minimized.With the emergence of the stock market, a lot of prediction techniques aredeveloped. Most of traditional prediction methods are based on long-term and largesample data, so they are suitable for forecasting medium or long-term stock. But thestock market is a complex nonlinear dynamic system, so the traditional predictionmethods are not suitable for short-term prediction of the stock market. The neuralnetwork which has the capabilities of non-linear approach and self-learning is verysuitable for solving some problems in the stock market. In recent years, the research ofapplication of artificial neural network to predict the stock market has become the hottopics in the study of scholars at home and abroad. And good results are achieved.In this paper, BP neural network which is a kind of widely used algorithms ischosen to forecast the stock price. But the BP algorithm is slow convergence speed andeasy into the local minimum points etc. To overcome these shortcomings, geneticalgorithm (GA) is used to optimize BP neural network.In this paper, two experiments are done: one is about BP neural network; the other is about BP neural network based on Genetic algorithm. The experimental resultssuggest that using genetic algorithm to optimize BP neural network can prevent the BPneural network into local minimum point and increase the convergent speed andimprove predicting accuracy. So the prediction method of stock price which applies BPneural network based on genetic algorithm works.