【摘要】 将优化BP神经网络应用于森林病虫害———马尾松毛虫的预测。在马尾松毛虫预测预报中,建立一个根据以往积累的数据,预测下一、二代直至下一年全年的马尾松毛虫有虫面积、发生面积、虫口密度、虫株率的系统。经实测检验,面积预测平均误差为6.86%,虫口密度预测平均误差为10.93%,虫株率预测平均误差为10.13%,说明该系统具有较高的可靠性,可应用于实际预测。更多还原

【Abstract】 This paper uses the BP network to forecast the occurring quantity of the plant diseases and insect pets.The forecast system of the Dendrolimus punctatus Walker is established.It includes forecast area of insect,insect density and percentage of insect′s tree for the next year′s every generation,according to the past data.The experiment shows that the system is advanced in credibility.更多还原