Research on the prediction method of grain yield basing on the BP network in Jilin province

Aiming at solving the problems of poor accuracy and large fluctuations in the grain yield prediction, the paper selects food production data of Jilin province in 1970-2011 as the research object, and takes 7 factors which influence agricultural production as the impact factors. The research adopts 2 prediction methods - regression analysis and the BP neural network analysis respectively, sets up the prediction models and makes the comparative analysis to the varies of prediction yield and actual production. The final end shows that the prediction mean accuracy of regression analysis is 86.9%, the prediction mean accuracy of the BP neural network analysis is 91.4%, the BP neural network is more suitable for grain yield prediction in Jilin province.