摘要
通过对1990～2007年河南粮食产量的分析,在影响粮食产量的诸多因素中选出农业生产条件等8个主要影响因素。基于粮食生产系统的复杂性,建立偏最小二乘回归与BP神经网络耦合模型。偏最小二乘法通过对自变量中的信息进行组合和提取,有效克服变量之间的多重相关性问题,实现了对高维数据的降维处理,同时降低了神经网络的输入维数,提高了网络的学习效率和稳健性,从而充分利用了2类现代建模方法的优点。结果表明,偏最小二乘神经网络耦合模型研究河南粮食产量的拟合精度和预测精度都比较理想。
The grain yield of Henan Province was analyzed from 1990 to 2007. There were many complex factors affecting the grain yield. So a composition model was proposed by combining neural network model with the partial least square method. The factors were analyzed by means of partial least square method to find the most important components, so that the problem of multi-correlation among variables could be solved, and the amount of input dimensions of the neural network could be reduced. When the neural network was applied, it could solve the non-linear problem and improve the expression ability of the model. The results showed that the proposed model had higher fitting accuracy and prediction accuracy.
出处
《安徽农业科学》
CAS
北大核心
2009年第28期13971-13973,共3页
Journal of Anhui Agricultural Sciences
关键词
农业生产条件
粮食
产量
偏最小二乘法
BP神经网络
Conditions of agriculture
Grain
Yield
Partial least square method
BP neural network
