应用二元联合熵函数,仿照一元最大熵函数的推导过程,构建测树因子二元最大熵概率密度函数,指出该函数实际上是二元多参数指数族分布,其幂是二维连续函数空间基的线性组合;一元与二元最大熵函数的构建可以得出:这种构建模型的方法可以推广到测树因子二元以上概率分布的情形;对二元最大熵函数与二元Weibull分布模型做对比分析,并指出前者具有更广的适应性;对已有二元函数如二元SBB函数与二元Beta函数做了概述,介绍SBB函数初值选取方法,并指出二元SBB函数能反映2个随机变量的相关程度;用二元最大熵函数、二元SBB函数与二元Beta函数分别测量浙江省域尺度毛竹胸径、年龄联合分布信息,结果表明前2者的测量精度非常高,都适合于描述毛竹胸径、年龄联合分布规律,回归离差平方和、R2与柯尔莫哥洛夫检验统计量依次为9.97677e-05,0.9960,0.99983;0.00084,0.96400,0.97998;二元Beta函数测量精度最低,函数初值选取与变量区间变换还有待于进一步研究。
Probability distribution of key factors in forest mensuration was very important in the management practice. It is poor understand that main variables multi-distribution of forest mensuration in the forest ecology,the bivariate distribution model of main variables of forest mensuration was introduced in this paper. We build up a bivariate maximum entropy probability density function used bivariate unite entropy function,and simulating unitary maximum entropy function. This approach demonstrated it was bivariate and exponential distribution for many parameters. It will prompt through to combine with the radix of bivariate consecutive function dimension. We analyzed and contrasted a bivariate maximum entropy function and a bivariate Weibull distribution,and pointed out the former had wider flexibility. The bivariate SBB function and the bivariate Beta function were introduced and the selection of SBB function's initial value had been elaborated and also pointed out that the bivariate SBB function could reflect the two random variables's correlation;bivariate maximum entropy probability density function,bivariate SBB function and the bivariate Beta function respectively were used to measure the two-dimension information of diameter-age. The results indicated that the precision of the former two methods were very high,and the two methods could be suited to describe the joint distribution of the bamboo diameter-age,for bivariate maximum entropy: minimum variance was 9. 976 77e-05,R^2 = 0. 996 0,A. H. Колмогоров statistical quantity was 0. 999 83,for bivariate SBB function: minimum variance was 0. 000 84,R^2 = 0. 964 00,A. H. Колмогоров statistical quantity was 0. 979 98. the precision of the bivariate Beta function was lowest,the selection of its initial value and the variable range transformation were further researched.