Research On Method Of Grain Output Prediction

Hunger breeds discontentment,the food problem is an important problem which related to the development of human society since ancient times.China is a developing country that is of large population but less land,the economic development has brought prosperity to the people's lives,but also put pressure on China's food supply.The supply and demand of food is closely linked with the issue of food security.The food security,as an important component of national security,which is not only an economic problem,but also related to social stability and the long-term development of the country,so it has become the focus of attention and study from all sectors of society to predict the grain yield effectively and ensure the food security in China.After in-depth understanding of the status of China's food,it mainly focusing on prediction work of grain production in this paper.Firstly,the output data of different time interval is predicted by using the improved ARIMA model and the traditional model which often used in time series analysis,experimental results validate that the prediction results will be more accurate when choose the modified model along with the selected time interval is longer.Secondly,the joint dynamic prediction model is proposed.Secondly,a dynamic prediction model of long-term grain output is realized with joint multivariate regression model,of which the correlation between the grain yield and its influence factors is analyzed firstly,then the influence factor in the model is predicted by using the improved ARIMA model.Finally,in order to effectively solve the nonlinear problem and parameter optimization with grain output data,the application principle of least squares support vector machine and the characteristics of global optimization of particle swarm algorithm is analyzed and studied based on statistical learning theory,then least squares support vector machine model optimized by the particle swarm algorithm(PSO-LSSVM)is proposed,which has a faster solution due to the effective combination of them,moreover it can guarantee the global optimal parameter selection,and the model with smoothing procedure can achieve higher-superiority in forecast precision compared with other models.It can be concluded through the analysis of the prediction results: the short-term and longterm prediction of grain output can be realized by using the improved ARIMA model and the combined dynamic model,moreover,the prediction accuracy of the latter is higher than that of the traditional model and grey model.Least squares support vector machine model which optimized by the particle swarm optimization algorithm can better solve the nonlinear to problem and achieve short-term grain output prediction.