基于BP神经网络的我国农民收入预测模型

依据1978~2008年影响我国农民收入因素的相关数据,选取从事农业的人口、第一产业产值、乡村就业人员数等13个指标,依据标准化方法和BP神经网络方法,建立了关于农民收入的人工神经网络模型,并进行具体分析。结果表明,模拟值与真实值吻合较好,改进BP算法的神经网络模型预测精度高,收敛速度快,具有良好的泛化能力。在此基础上,提出了增加农民收入的建议：一是推进城镇化进程;二是发展农村中小企业;三是鼓励集约经营;四是加强农村基础设施建设和农业科技投入。

A Prediction Model of Farmers' Income in China Based on BP Neural Network

According to the related data affecting the farmers' income in China in the years 1978-2008,a total of 13 indices are selected,such as agricultural population,output value of primary industry,and rural employees.According to standardized method and BP neural network method,the farmers' income and the artificial neural network model are established and analyzed.Results show that the simulation value agrees well with the real value;the neural network model with improved BP algorithm has high prediction accuracy,rapid convergence rate and good generalization ability.Finally,suggestions are put forward to increase the farmers' income,such as promoting the process of urbanization,developing small and medium-sized enterprises in rural areas,encouraging intensive operation,and strengthening the rural infrastructure and agricultural science and technology input.