摘要
为了在生物量建模过程中得到回归模型的最优估计,针对回归模型存在的异方差性,提出用怀特检验方法来定量分析模型,再结合残差分布图来辅助判断,并应用怀特检验方法和残差分布图实例分析了普通最小二乘法拟合栎类生物量模型结果存在异方差性,建议采用加权最小二乘法拟合栎类生物量模型。
In order to have the optimal estimation during regression model for biomass, White test was recommended to quantitative analyze model, associated by residuals distribution. Application of White test and residuals was conducted on analysis of biomass model for Quercus L. fitted by ordinary least squares, the result showed that there had heteroskedasticity. Therefore, weighted least squares were recommended to fit biomass model for Quercus L.
出处
《浙江林业科技》
北大核心
2012年第3期43-45,共3页
Journal of Zhejiang Forestry Science and Technology
