论文部分内容阅读
基于黑龙江省孟家岗林场60株人工红松955个标准枝数据,采用线性混合效应模型理论和方法,考虑树木效应,利用SAS软件中的MIXED模块拟合红松人工林一级枝条各因子(基径、枝长、着枝角度)的预测模型.结果表明:通过选择合适的随机参数和方差协方差结构能够提高模型的拟合精度;把相关性结构包括复合对称结构CS、一阶自回归结构AR(1)及一阶自回归与滑动平均结构ARMA(1,1)加入到一级枝条大小最优混合模型中,AR(1)可显著提高枝条基径和角度混合模型的拟合精度,但3种结构均不能提高枝条角度混合模型的精度.为了描述混合模型构建过程中产生的异方差现象,把CF1和CF2函数加入到枝条混合模型中,CF1函数显著提高了枝条角度混合模型的拟合效果,CF2函数显著提高了枝条基径和长度混合模型拟合效果.模型检验结果表明:对于红松人工林一级枝条大小预测模型,混合效应模型的估计精度比传统回归模型估计精度明显提高.
Based on the data of 955 standard branches of 60 artificial Korean pine trees in Mengjiagang Forest Farm, Heilongjiang Province, the linear mixed effect model theory and method were used to consider the tree effect. The MIXED module in SAS software was used to fit the factors , Branch length, branch angle) .The results show that the fitting accuracy of the model can be improved by choosing appropriate random parameters and variance covariance structure, and the correlation structure includes composite symmetry structure CS, first-order autoregressive structure AR (1) and ARMA (1,1), a first-order autoregressive and moving average structure, are added to the first-order optimal branch size model. AR (1) can significantly improve the fitting accuracy of the mixed root- All three kinds of structures can not improve the accuracy of the mixed model of branch angle.In order to describe the heteroscedasticity phenomenon generated in the hybrid model construction, CF1 and CF2 functions are added to the branch mixed model, CF1 function significantly improves the fitting of the branch angle mixed model Effect and CF2 function significantly improved the fitting effect of mixed model of branch diameter and length.The model test results showed that for the first-order branch size prediction model of Pinus koraiensis plantation, the mixed effect The estimation accuracy of the model is obviously higher than that of the traditional regression model.