Research On China's Grain Consumption Structure Based On BP Network Optimization

From ancient times until now,food problem has always been an important reserve resource related to national economy and people's livelihood,and the study of food consumption demand has always been the top priority.Under the background of rapid economic development and continuous improvement of residents' income level,people's demand for food began to change,showing a diversified trend of development,specifically reflected on the changes in the structure of food consumption,which will affect social stability,on the other hand,will have an impact on the physical quality of the whole people.Therefore,the scientific prediction of China's grain consumption demand can provide a reference for adjusting grain production structure and formulating grain consumption policy.On the premise of fully understanding the current situation,influencing factors and forecasting methods of grain consumption structure in China,this paper constructs time series forecasting model and multi-influencing factor forecasting model respectively.Considering the non-linear characteristics of grain consumption data,and BP network is suitable for most complex non-linear problems,so a forecasting model based on BP network is constructed.The main work and creative ideas of this paper are as follows:1.Re-evaluate the relevant factors which affect China's food consumption,especially some new factors that are forming,and make a quantitative analysis of them.2.From the point of analyzing the changing trend of grain consumption demand with time trend,a grey BP network model based on wavelet denoising is constructed.Firstly,the original data is denoised by wavelet transform,then the denoised data is used as initial value sequence for GM(1,1)prediction,and finally the BP network is used for residual correction and predicted the consumption of rations and fodder in China in the next three years,and its accuracy is higher than that of the traditional grey model.3.By analyzing the quantitative relationship between many influencing factors and grain consumption demand,the MIV-IPSO-BP network model was constructed.Firstly,the main factors affecting grain consumption are screened out by means of average impact value method.Then,BP network model is constructed.In order to improve the prediction accuracy of the model,IPSO algorithm is used to optimize the connection weights of BP network to find the global optimal solution.The consumption of rations and fodder grains in China in the next four years is predicted.It can be seen that the prediction accuracy of the model has been improved after variable selection and algorithm optimization.The simulation results show that the Grey BP network prediction model and MIV-IPSO-BP network prediction model after wavelet denoising are feasible in predicting China's grain consumption,which shows that the proposed method can effectively predict China's grain consumption.