论文部分内容阅读
通过在具有动态反馈机制的Elman神经网络的基础上引入时间收益因素,提出了一种改进的Elman神经网络模型并将其用于对股票的综合指数进行预测,进而求其收益率。实验模拟结果表明:将改进的Elman模型用于股市投资是可行的,有效的,具有一定的应用潜能,该模型不仅可以明显提高网络的预测精度,达到快速收敛,而且还能够明显提高股民投资的利润率,实现较大幅度地获得收益的目的。
Based on the Elman neural network with dynamic feedback mechanism, this paper introduces a time earnings factor, and proposes an improved Elman neural network model which can be used to predict the composite index of stocks and then calculate the return rate. Experimental results show that the improved Elman model is feasible, effective and has potential application in the stock market. This model can not only significantly improve the prediction accuracy of the network and achieve fast convergence, but also significantly improve the investment of the investors Profit margin, to achieve more substantial gains.