关键词: 农产品, 价格预测模型, 卷积神经网络, 门控循环单元, 短期预测

Abstract: In view of the fact that the existing forecast model could not accurately and rapidly predict the price of agricultural products in the large data environment, a new forecast model for the price of agricultural products based on Convolutional Neural Network（CNN）and Gated Recurrent Unit （GRU）was presented. Local features were obtained by using CNN, and time series dependence of data was attained by using GRU. Then, the features gained by the two were connected, and the predictive output was obtained through the decoder. The superiority of the model was verified by the comparative test with the traditional single model. The result showed that, compared with the traditional prediction model, the model built in this study could effectively conduct short-term prediction, and had a certain practical value for predicting the price of agricultural products.

Key words: agricultural products, price prediction model, convolutional neural network, gated recurrent unit, short-term forecast