摘要
根据逐步回归法分别筛选出与松墨天牛(Monochamus alternatus Hope)、短角幽天牛(Spondylis bupres-toides(L.))、立毛角胫象(Shirahoshizo erectus Chen)、松瘤象(Hyposipalus gigas Linnaeus)4种马尾松(Pinus massoni-ana)主要蛀干害虫种群动态变化密切相关的因子,然后用似乎不相关线性回归方法分析了各蛀干害虫林间诱捕序列的动态变化及其相互联系。结果表明:在特定情形下似乎不相关线性回归估计能够更好地反映马尾松4种主要蛀干害虫林间种群的相互关系,其回归系数的估计精度优于普通最小二乘法。
Factors having close relationships with population dynamics of 4 main borers in Piaus massoniana were selected by the method of stepwise regression. Then the population dynamics and relationships of trapping series of Monochamus alternatus Hope, Spondylis buprestoides L. , Shirahoshizo erectus Chen, and Hyposipalus gigas Linnaeus were analyzed using seeming- ly unrelated linear regression method. Results show that, under specified conditions, the seemingly unrelated linear regression method can effectively reflect the relationships between different populations of the four main borers, and its forecast accuracy is higher compared with the ordinary least squares method.
出处
《东北林业大学学报》
CAS
CSCD
北大核心
2012年第11期135-138,共4页
Journal of Northeast Forestry University
关键词
蛀干类害虫
林间诱捕序列
似乎不相关
种群动态
普通最小二乘法
马尾松
Stem borers
Trapping sequence in forest
Seemingly unrelated regression
Population dynamics
Ordinary least squares
Pinus massoniana