The stock index forecast based on dynamic recurrent neural network trained with GA

The stock index forecast based on dynamic recurrent neural network trained with GA Fang Yixian1，∗ Wang Baowen2 Wang Yongmao1 1 Science of college,Yanshan University, Qinhuangdao Hebei Province, China,066004 2 Information of college, Yanshan University,Qinhuangdao Hebei Province, China,066004 Abstract：In order to forecast the stock market more accurately, according to the dynamic property for the stock market, propose the real time modeling forecast via dynamic recurrent neural network and use GA to study online, then it improves the network performance and better describes the dynamic characteristic of stock market. By forecasting Shanghai negotiable securities index, it shows better validity. Keyword: Dynamic property Neural network Stock index Genetic algorithm 1 Introduction The stock market is a very complicated nonlinear dynamic system, it has both the high income and high risk properties. So the forecast of stock market trend has been always paid attention to by stockholders and invest organization. However, because of the high nonlinearity of the stock market, it is difficult to reveal the inside law by the traditional forecast methods, so we are not satisfied with many of the applied effect for forecast analysis method [1]. In recent years, the rapid development of computer and artificial intelligence technology provide many new technology methods for the modeling and forecast of the stock market [1]. The neural network gains wide application in the aspect of nonlinear forecast due to its broad adaptive and learning ability. There is many neural networks applied in forecasting stock price, at present, the most widely used neural network is BP NN, but BPNN has many shortcomings such as the slow learning rate, large calculate amount, easy to get into local minimum and bigger randomicity and so on. This affects the predicted results of the stock price. RBF neural networks is also a very popular method to predict stock market, this network has better calculational and spreading abilities, it also has stronger nonlinear mapped ability[2]. But the stock market is not only with nonlinearity but also chaos, and it is a dynamic system related to time (time is a independent variable). Therefore the network for predicting itself is a dynamic system, it requires us to introduce short-time memory function in network. According to short-time the dynamic neural network can be divided into time delay NN and local or whole feedback NN. This paper uses an improved Elman dynamic recurrent neural network model[3-4] to carry modeling forecast for stock index, this neural network has better effect in dealing with time varying information aspect, the results show that using it to predict time varying stock index will have satisfactory effect. 2 Neural network predict approach for stock price ∗ E-mail: jiahao218@126.com, Tel: 0335-8515601 319 Let { }( 1,2, Xt t N = L ) be the data series of time-varying volatility for stock price, the stock predict is to set up a modelG , and describe the relation between j X and prevenient price ( ) 1 2 , , , j j j k X X X − − − …

Content truncated to 500 words