林木生物量参数的非线性最小二乘解法研究-吉林农业大学学报.PDF

DOI :10.13327/j.j jlau.2006.03.007 　28 3 　　　　　　　 Vol.28 No.3　 　2006 6 Journal 　of 　Jilin　Agricultural　University June 2006 　 1, 2 2 王仲锋 , 冯仲科 (1.北京林业大学资源与环境学院, 北京 100083;2.长春工程学院勘查与测绘学院, 长春 130021) 　:为了提高林木生物量参数的计算精度, 根据林木相对生长公式(幂函数), 阐 用非线性最小二乘法 求解生物量参数估值的原理、方法和过程;给出计算参数估值之方差与协方差的公式;揭示幂函数对数线性化 回归的本质。 通过理论和实例分析, 将非线性最小二乘 (NLS)法与常用的对数线性化最小二乘(LLLS)法进行 比较。结果为用NLS 求解的林木生物量参数与用LLLS 求解的林木生物量参数相比有明显的差异;用NLS 求 解的林木生物量参数的精度高于用LLLS 求解的林木生物量参数的精度;用NLS 回归的林木生物量方程比用 LLLS 回归的林木生物量方程具有更好的相容性;LLLS 法具有异方差性, 不满足标准线性回归的基本假设。 :森林生物量;非线性最小二乘法;相对生长公式;方差与协方差 :S718.5 　　　:A　　　:1000-5684(2006)03-0261-04 On Estimating the Parameters of Tree Biomass Using Nonlinear LS WA NG Zhong-feng1, 2, FENG Zhong-ke1 (1.College of Resources and Environment, Bejiing Forestry University , Beji ing 100083, China ; 2.School of Prospecting and Surveying, Changchun Institute of echnology , Changchun 130021, China ) Abstract:To improve the precision of tree biomass parameter calculation, this paper e plicates the prin- ciple, method and process to estimate the parameters of tree biomass by the nonlinear least square accord- ing to the formula of allometry method (power function).It gives the formula to calculate the variance and covariance estimated parameters, and reveals the essence of logarithmic linearization regression of power function.With the theoretical and practical data analysis, it compares the nonlinear least square method (NLS)with the logarithmic linearization least square method (LLLS).The concrete conclusions in the pa- per are that tree biomass par