摘要
应用最大熵原理构造了测树因子概率分布的统一模型,这样构造的模型具有明确的解析表达式,并能克服常用方法无法解释测树因子服从某种概率分布的真正原因,从而为测树因子统计分布建模提供了一种有效方法。使用1-3阶样本矩、1-4阶样本矩与1-5阶样本矩,用所构建的概率分布统一模型分别对浙江省域毛竹胸径分布分别作了仿真试验,结果表明当采用1-4阶样本矩时,仿真效果最好,而且比通过假设检验的Weibull分布仿真结果理想:(1)图形非常相似,对实测数据都能很好的模拟;(2)最大熵法的离差平方和为0.00018,Weibull分布的为0.00045[1]。由于各种系统与非系统的原因,都会影响测量结果的准确性,对所构建的模型作了不确定度评定,表明结果具有很大的可靠性,测量结果的估计:7.85100,测量结果的标准不确定度:1.82710,置信概率:0.96020。
This paper used a maximum entropy theory to establish a general probability distribution model for tree measurement parameters.This model has an explicit explanatory expression and overcomes some problems occurred in the traditional methods that the reasons why the tree measurement parameters obeyed certain probability distribution cannot effectively be explained.Therefore,this model provides a new way to establish the statistical distribution for tree measurement parameters.The established general probability distribution model was used to simulate the Moso bamboo′s diameter distribution in Zhejiang province based on 1-3,1-4,and 1-5 stage sample moments,respectively.The results indicated that using 1-4 stage sample moment provided the best simulation performance,and provided even better effects than that using Weibull distribution.Both maximum entropy theory and Weibull distribution have similar features that can effectively simulate the reference data.The sum of square deviation is 0.00018 based on maximum entropy theory and 0.00045 based on Weibull distribution.Because different system and non-system factors can affect the reliability of the estimates,the established models was used to evaluate the measurement uncertainty,indicating the reliable results with estimates of 7.85100,standard uncertainty of 1.8271,and confidence probability of 0.9602.
出处
《生态学报》
CAS
CSCD
北大核心
2009年第1期86-91,共6页
Acta Ecologica Sinica
基金
国家自然基金资助项目(30771725)
国家自然基金资助项目(30700638)
关键词
最大熵原理
概率分布模型
WEIBULL分布
测量不确定度
maximum entropy theory
probability distribution model
Weibull distribution
measurement uncertainty
