综合改进BP神经网络算法在股价预测中的应用

BP神经网络用于股市等非线性系统的预测具有非常强大的能力。针对标准BP神经网络算法存在收敛速度慢和易陷入局部极小点问题,提出附加动量法和动态调整学习率法相结合的综合改进BP算法。建立了基于改进算法的股票价格预测系统模型,并对青岛海尔股票价格进行了预测。仿真结果表明,相对于标准BP算法,综合改进的BP算法能明显加快网络的收敛时间,用于股价短期预测是有效可行的,并且能够达到较高的准确性,具有一定的实际应用价值。

Application of Integrative Improved BP Neural Network Algorithm in Stock Price Forecast

BP neural network to forecast the stock market and other non-linear system has a very strong capacity.In view of the standard BP neural network algorithm exists slow convergence and easy to fall into local minimum,this paper presents an integrative improved BP algorithm which combines additional momentum and dynamic adjustment learning rate.A system model based on improved algorithm for forecast stock price is studied and stock price of Qingdao Haier is forecasted in this paper.The simulation result shows t...