欢迎访问《中国农学通报》，
中国农学通报

中国农学通报 ›› 2014, Vol. 30 ›› Issue (28): 72-75.doi: 10.11924/j.issn.1000-6850.2014-1535

• • 上一篇    下一篇

基于人工神经网络的落叶松毛虫发生量预测模型的研究

杨淑香,赵慧颖   

  1. 内蒙古呼伦贝尔市气象局,黑龙江省气象科学研究所
  • 收稿日期:2014-05-29 修回日期:2014-05-29 接受日期:2014-06-23 出版日期:2014-10-15 发布日期:2014-10-15
  • 通讯作者: 杨淑香
  • 基金资助:
    内蒙古自治区气象局科技创新基金项目(nmqxkjcx201407)

A Study on the Forecast Model of Dendrolimus superans Butler Occurrence Based on Artificial Neural Network

  • Received:2014-05-29 Revised:2014-05-29 Accepted:2014-06-23 Online:2014-10-15 Published:2014-10-15

摘要： 运用人工神经网络的原理和算法，根据相关系数法和逐步回归法选取了蒸发量、气温、风速、相对湿度等气象因子作为预报因子，建立了内蒙古东部地区的鄂伦春自治旗落叶松毛虫的发生面积及虫口密度与气象因子之间的BP网络模型。结果表明：所建立的模型具有较高的预测效果。通过逐步回归筛选出的预报因子，与事实吻合，选取合理。误差较小，控制在0.1%~25.0%之间。可以作为病虫害综合防治的依据。

关键词: 聚类分析, 聚类分析

Abstract: The principle and algorithm of artificial neural network were used to select some meteorological factors to established the BP network model of occurrence area and population density of Dendrolimus superans and meteorological factors in Elunchun Zizhiqi of eastern Inner Mongolia, based on the correlation coefficient method and the stepwise regression method for selecting the evaporation, temperature, wind speed, relative humidity and others as the forecast factors. The results showed that the established model had higher prediction effect. The reasonable forecast factors matched with the facts. The error was small, which was controlled between 0.1%-25.0%. The model could be used as the basis for integrated pest management.