对权函数法求概率分布参数的讨论

Discussion on determination of probability distribution parameters using weighting function arithmetic

本文在分析以往研究成果的基础上 ,提出了用单权函数数值积分逼近算法 (即采用一个权函数 ,结合数值积分法 ) ,可以同时计算出样本的变差系数和偏态系数 .给出了两种单权函数的详细求解过程 ,并对“数值积分双权函数法”进行了改进 ,用数学计算软件Mathcad 7 0设计的计算程序 ,作了实例计算 .并且与相同样本条件下“数值积分双权函数法”和“数值积分矩法”的计算结果进行了对比分析 .

A numerical integral approach arithmetic using weighting function is proposed.It can solve the coefficients of sampling variation and skewness simultaneously.The solution procedures of exponential and variation single weighting functions are introduced in detail.The numerical integral arithmetic of double weighting function is also improved.An example solved by using Mathcad 7\^0 is given for demonstration.The calculation result is compared with those using arithmetic of double weighting function and moment method of numerical integral.