Multi-factor Corn Yield Prediction Based On Gray Neural Network

For each country,ensuring steady crop production is directly related to the stability of the country and the long-term economic development,China is not only a large agricultural country,but also a large country with a large population.The abundant crop yield always affects all walks of life,affects everyone,corn as an important food and industrial raw materials,has become an important indicator of China's industrial and agricultural development,China's crop safety and future development also has a direct impact.Therefore,it is very necessary to explore and study the future trend of corn production in China.This paper takes the total annual output of corn and the annual unit output data of 33 years between 1985 and 2017 in North China as the research object.Using the GM(1,1)model and BP neural network,the combined multi-factor corn yield prediction model is established and the model is trained to optimize,so as to improve the prediction accuracy of the model.The main work and conclusions of the thesis are as follows:(1)To establish a forecast model of annual corn production and unit yield in North China based on gray GM(1,1).Taking into account the characteristics of corn yield over time,a dynamic corn yield prediction model is established by adjusting the steps of the gray GM model.(2)The main control factors affecting the total annual output of corn were studied and analyzed.The degree of correlation between 12 influencing factors and corn yield is calculated and sorted by the gray correlation analysis method in gray theory.The results of the analysis of the main factors put forward effective suggestions and reasonable measures.(3)The constraint indicators after processing are used as input variables of BP neural network prediction algorithm,a corn yield prediction model is established,and MATLAB is used to program the prediction of annual corn production and yield in 2013-201 7.(4)Build a combined prediction model of gray neural network.The parameters of the gray differential equation correspond to the weight and threshold of the neural network,and the optimized parameters are determined by training the neural network to make it steady.The training results of BP neural network are then fitted with the results obtained by the gray GM model(1.1).Finally,the prediction results of the three models of gray GM(1.1),BP neural network and gray neural network model are compared and analyzed.The prediction results show that the GM(1,1)model can realize the prediction of annual corn production based on time series,and its error rate is about 5.16%,and the model accuracy is second-level.Under the premise of taking into account the complexity of corn yield factors,the prediction method based on BP neural network model is proposed,and the error of total yield forecast and yield prediction is 1.68%and 2.48%,and the accuracy is improved compared with the traditional gray model.The absolute percentage error of corn yield prediction based on the combination-based gray neural network model decreased to 1.09%and 1.75%,and the prediction accuracy was significantly improved compared with the gray theory model and The BP neural network model.Above all,the characteristics and results of the three models are combined,and the model algorithm established by the combined model is selected to predict the yield of corn in North China.