论文部分内容阅读
利用南方杉木(Cunninghamia lanceolata)的240株建模样本数据,通过建立一元、二元立木材积和地上生物量回归模型,对特定权函数、通用权函数及其拓展的4个权函数的加权回归估计结果进行了对比分析。结果表明:针对某一组具体建模数据而言,采用特定权函数进行加权回归估计的做法是合适的;采用通用权函数进行加权回归估计,对不同的建模数据,其消除异方差的效果并不完全相同;为了使通用权函数具有更广泛的适应性,建议将1/f(x)2调整为1/f(x)n。
Based on the data from 240 modeling samples of Cunninghamia lanceolata, we established a one-way, two-component tree volume and aboveground biomass regression model to estimate the weighted function, the universal weight function and the four weight functions The results of a comparative analysis. The results show that it is appropriate to use a specific weight function to estimate the weighted regression for a specific set of modeling data. Weighted regression estimation using a universal weight function can eliminate the effect of heteroscedasticity for different modeling data Are not exactly the same; in order to make the universal weight function more widely adaptable, it is suggested to adjust 1 / f (x) 2 to 1 / f (x) n.