神经网络用于股市预测的一种新方法

A New Approach for Stock-market Prediction by Application of Neural Network

采用并联的神经网络结构,将原始数据和技术指标值分别输入两个不同的子网络,在输出端将两者的输出综合起来得到一个新的输出,以此提高预测精度.最后以上证指数为例,通过仿真验证这种方法的可行性.

By the adoption of a parallel structure of neural networks,the initial data and the data index values are inputted into two different sub-networks. At the output terminal,the outputs of the former two are synthesized to obtain a new output,by way of which,the precision of prediction is improved. Finally,the Index of Shanghai Stock Exchange is taken as an example to verify the feasibility of this approach by way of simulation.