主成分分析在小麦条锈病预测中的应用

影响因子过多,不但会引起数据"维爆炸",而且会降低预测精确度,为了解决该问题建立了基于主成分分析的BP神经网络模型.首先阐述了主成分分析法的原理与步骤,然后分析了陕西省汉中地区1974～1997年小麦条锈病病情及相关气象资料,得出影响其流行的主要成分,最后利用得到的主要成分作为BP神经网络的输入,对汉中地区1994～1997年小麦条锈病流行情况进行预测,并与未进行主分分析而建立的全要素BP网络模型进行比较.实验结果表明,该模型可以快速准确地预测小麦条锈病的流行程度,有效地减少小麦产量损失.

Application of principal component analysis of wheat stripe rust

To deal with the too much data will be handled before building the prediction model,even more the precision of the prediction model will be lowed due to lots of factors,the BP neural network model based on principal component analysis(PCA) is proposed.Firstly,the principal and the procedures of PCA are introduced.Secondly,the original data recorded by Han Zhong area between 1974 and 1997 is analyzed by PCA and the main factors which lead to the wheat stripe rust are gotten.Finally,the obtained key component...