神经网络在松墨天牛发生量预报中的应用  

Forecast of the occurrence quantity of Monochamus alternatus by using neural network

作　　者：陈顺立[1] 张华峰[1] 张潮巨[2] 谢峥 

机构地区：[1]福建农林大学林学院,福建福州350002 [2]福建省光泽华侨国有林场,福建光泽354100 [3]福建省古田县林业局,福建古田352200

出　　处：《福建林学院学报》2006年第1期6-9,共4页Journal of Fujian College of Forestry

基　　金：福建省科学技术厅重大科技项目(2000-Z-028)

摘　　要：运用神经网络的基本原理,结合闽北地区气象和环境因子及松墨天牛发生量的实测数据,建立神经网络模型.结果表明:所建立的BP神经网络模型,具有满意的拟合精度和预测精度.2个预留调查点的平均预测精度达97.2%,预测准确率为100%.Based on the theory and methods of neural network, climatic data, geographical data and the occurrence quantity in north of Fujian province, an occurrence quantity forecast of Monochamus alternatus Hope was established. The results showed that the established neural network model had both satisfying fitting and forecasting precision. The mean forecast precision for occurrence quantity was 97.2% in two survey plots, and forecast accuracy was 100%.

关 键 词：松墨天牛 神经网络 发生量 预测预报 

分 类 号：S763.3[农业科学—森林保护学]