基于小波神经网络的麦蚜发生量预测研究

Study on the Prediction of Occurrence Quantity of Wheat Aphids Based on Wavelet Neural Network

以1979—2014年麦蚜历年发生量及气象数据为基础,采用逐步回归法筛选出对麦蚜发生量影响最大的气象因子,以筛选出的气象因子和1979—2009年麦蚜发生量作为训练集建立了小波神经网络预报模型,用该模型对2010—2014年麦蚜发生量进行预测,结果显示,小波神经网络预报模型的预报结果较准确。

Baesd on the Wheat Aphids and meteorological data of 1979-2014, meteorological factors that influenced the maximum amount of wheat aphids occurred were screened by stepwise regression method, and then taking the meteorological factors and 1979- 2009 which were screened or years Aphids occurrence amount as the training set, the wavelet neural network of forecast model was built, using the model, occurrence quantity of 2010-2014 years of wheat aphids were predicted, the results showed that the pre- diction result of wavelet neural network prediction model was accurate.