基于Elman递归神经网络的股价的短期预测

Short Term Prediction of Stock Price Based on Elman Recursive Neural Network

Elman神经网络是一种典型的局部递归神经网络,非常适合用于如金融时间序列这样复杂的非线性动力学系统的预测中。用美菱电器(股票代码:000521)和上海电力(股票代码:600021)的280天的实际开盘价作为时间序列预测的样本,用Elman递归神经网络方法建立股票价格预测模型。通过Matlab软件对其预测过程进行仿真实验,验证了Elman神经网络建立的股票开盘价短期预测模型具有收敛速度快、预测精度高等优点。

Elman neural network is a typical local recurrent neural network, which is very suitable for the prediction of complex nonlinear dynamic system such as financial time series. This paper used Mei ling electronics （Stock Code： 000521） and Shanghai electric power（Stock Code：600021 ） for 280 days of the actual opening price as time series prediction samples. It used the Elman recursive neural network method to establish the stock price forecasting model and simulate prediction process through the Matlab software. In this way, the stock opening price short-term forecasting model had the advantages of fast convergence and high precision.