Price Prediction of Traditional Chinese Medicine Based on ARIMA and Improved Elman Neural Network

Abstract

The price’s change of traditional Chinese medicine contains linear, non-linear and other miscellaneous factors. It is difficult for people to use a separate model such as neural network model to judge its price trend. Based on the background, a combined forecasting model is proposed in this paper, it consists of Autoregressive Integrated Moving Average model and Elman neural network which is improved by correlation analysis. The combined forecasting model can use its two algorithm model to deal with the linear and nonlinear factors. Meanwhile, the innovation of this paper is using correlation analysis to import extra additional parameters for the neural network, which can increase its accuracy. A large number of traditional Chinese medicine’s price data was collected to be training samples, the final results show that the combined forecasting model has advantages over stability and accuracy than ARIMA or Elman neural network.