Corn is one of the three main crops with the widest planting area,the highest yield and the most eaten in China.Mastering the technology of scientific prediction of corn yield can provide technical support for agricultural planting planning,grain storage and processing,as well as market regulation.Taking into account meteorological factors and soil factors,this paper establishes BP neural network model,RBF radial basis neural network model,and GBDT gradient lifting decision tree model;then,the paper makes a regression analysis of corn yield in various counties and cities of Jilin Province,and a comparative analysis of their errors.In the experimental results,the fitting degree between the predicted yield and the real yield of GBDT model is high,R2 is up to 0.92,which can show a good effect in the prediction of corn yield in various counties and cities of Jilin Province.The results show that the model is feasible to predict the corn yield of 40 counties and cities in Jilin Province,and the data are easy to obtain,thereby can guide the agricultural departments of the government to formulate relevant policies and guidelines to guide production.
国家科技期刊平台
NSTL
万方数据
