The Application And Empirical Analysis Of BP Neural Network Model For Stock Market Forecasting

Nowadays,with the development of the economy and the increase of investment awareness,stock investment has become an important way for people to invest today,so the prediction of stock prices has also become a matter of concern to investors and researchers.The results show that the stock market is a nonlinear dynamic system.Using the traditional time series model forecasts the stock price,due to the weak performance of its nonlinear mapping and the difficulty in determining the characteristics of the appropriate model structure,the stock price prediction results are generally not satisfactory.Therefore,how to establish a stock price forecasting model with high calculation speed and accuracy is the focus of current financial investors' concern and research.The BP neural network(Back Propagation)proposed in this paper,as a modern intelligent information processing method,has the characteristics of self-adaptive data learning,nonlinear mapping and so on,and is suitable for dealing with complex nonlinear problems such as stock price.The BP network model will learn the inherent laws of stock price development by learning past history data,and store it in the model's specific weights and thresholds,and finally be used to predict future trends.In this paper,the existing BP neural network model has the problems of slow learning and low accuracy of prediction results.By re-selecting the activation function of the neuron,the weights and scaling factors of the neuron transfer function in the output layer and the hidden layer are improved.The displacement parameters are adjusted to reduce the number of hidden layer nodes and speed up the convergence of the network.Finally,based on the improved BP algorithm,a prediction model is built.Taking the actual stock as an example,the future trend of its stock price is forecasted and a better result is obtained.It is proved that the BP neural network model is in the stock market.The feasibility and effectiveness of price forecasting.