一种新的昆虫神经网络预测预报方法

A NEW PREDICATIVE NEURAL NETWORK METHOD FOR INSECT FORECAST

昆虫预测预报是植物保护中的一个困难问题．本文首先从系统的观点出发，分析了影响昆虫种群密度的主要因素；在研究传统时间序列模型共性的基础上，给出了广义时间序列模型；提出了一种新的昆虫预测模式—神经网络预测预报方法；用此方法对棉田害虫烟蓟马进行了预报，数值结果表明这是一种很好的预测预报方法，预测精度较其它方法都有很大提高．

The insect forecast is a difficult problem in plant protection. This paper has analysed the main factors which influence the insect density and made a new neural network predicative model. This method has a capacity for learning and can solve the forecasting problem which the traditional method could not. We have used this method to forecast the insect density in cotton field.The result shows that this method is better than other predicative methods and very useful in practice.