A Comparative Study Of China's Agricultural Product Output Prediction Models

In the existing research of China’s agricultural yield prediction,it has been applied to the prediction of various agricultural products in China and has achieved good prediction results based on the ARIMA model,the grey prediction model and the model average prediction method under the OPT criterion.In order to improve the prediction level of China’s agricultural yield,it is of great significance to compare the prediction errors of China’s agricultural yield based on different model prediction methods.This thesis makes an empirical study of the prediction of China’s agricultural yield.Through the ARIMA model,the grey prediction model and the model average under the OPT criteria,the three prediction methods for China’s total grain yield are used for the next three years’ short-term prediction and the next nine years’ medium-long-term prediction.Short-term prediction errors and medium-long-term prediction errors under the prediction method of grain yield are compared in different models.At the same time,by predicting the agricultural yield of corn,a specific agricultural product,we analyze the short-term and medium-long-term forecast results of corn yield under the three prediction methods,and determine whether the prediction results are consistent.The empirical analysis shows that for medium-long-term prediction and short-term prediction of China’s total grain yield,the model average method under the OPT criterion has the best forecast effect,while the average relative error of the medium-long-term and short-term forecasts of the ARIMA model are smaller than that of the grey prediction model.In the short-term prediction of corn yield,the average relative error of grey prediction and the model average method prediction under the OPT criterion are small,and the prediction effect of the model average method under the OPT criterion is better than that of the grey prediction,and the prediction effect of the ARIMA model on the corn yield is worst.Through medium-long-term predictions of corn yield,it is found that the three prediction methods have poor prediction results,but the prediction error of the model average method under the OPT criterion is lower than that of the grey prediction and ARIMA model prediction.Therefore,the short-term and medium-long-term optimal prediction models for total grain yield and corn yield are relatively consistent,and the prediction results are not consistent.