Export Volume Forecasting Model based on ARIMA-BP Neural Network

In view of the existing problems in the evaluation and prediction of the current data analysis, this paper carries out data mining and analysis according to the dynamic change trend of different indicators to be measured, so as to set three evaluation indicators according to the system for analysis, so as to make gray prediction of the change trend of the measured indicators, and carry out the overall evaluation analysis and prediction evaluation of the current three parts and other indicators. In this paper, the continuous values of the indicators to be measured were set up a database, and the ARIMA model, BP neural network model and their combination model were constructed to predict the change trend of the indicators, and the model with the highest degree of fit was obtained by comparing MAE, MAPE and RMSE three indicators. Thus, the influence trend and influence time of different influence factors are obtained, and the prediction results of the indicators to be measured are obtained by fitting the model.