P—Ⅲ型分布参数估计的模糊加权优化适线法2.docVIP

P—Ⅲ型分布参数估计的模糊加权优化适线法2行政论文范文大全 P—Ⅲ型分布参数估计的模糊加权优化适线法 3　各种参数估计方法的对比检验 　　3.1 计算结果 任何参数估计方法，即使在理论上正确，也必须通过大量适当的数据充分检验合格以后，才有应用的价值，本文采用刘光文教授［6］提出的与水文安全因数有关的绝对精度标准，即x1%和x0.1%(作为代表)的真误差不应超过8%—10%，用本文提出的模糊加权优化适线法和现行的几种参数估计方法，对大量的理想样本资料同时进行了对比计算，以下仅将与文献［6］中20组理想样本对 比的计算结果列于表1. 　　　　　　　　　　　表1 假想样本对各种p—ⅲ型参数估计方法的对比检验 样本序号* 样本容量(n) 项 目 假想理论值 (tv) 数值积分* 双权函数法 (nidwf) 模糊加权最小 平方准则 (fwlsc) 模糊加权绝对值 最小准则 (fwlavc) 1 19 cv cs x1% x0.1% 1 1/2 1 2.51128 3.26556 0.99344(-0.7%) 0.48154(-3.7%) 0.96948(-3.1%) 2.42987(-3.2%) 3.14018(-3.8% 0.97942(2.1%) 0.5046(+1.3%) 0.8573(14.3%) 2.4184(-1.2%) 3.18867(-2.4% 0.97942(2.1%) 0.5089(+1.8%) 0.7474(25.3%) 2.44141(2.4%) 3.11913(4.5%) 2 29 cv cs x1% x0.1% 1 1/2 1 2.51128 3.26556 0.99561(-0.4%) 0.48796(-2.4%) 0.98480(-1.5%) 2.45920(-2.1%) 3.18634(-2.4%) 0.98454(1.5%) 0.5097(+1.9%) 0.9041(-9.6%) 2.51049(0.0%) 3.24205(1.0%) 0.98454(1.5%) 0.4953(-0.9%) 0.6418(35.8%) 2.38095(5.2%) 2.99158(8.4%) 3 19 cv cs x1% x0.1% 1 1/2 3/2 2.66518 3.61676 0.99072(-0.9%) 0.47775(-4.5%) 1.47686(-1.5%) 2.56057(-3.9%) 3.45267(-4.5%) 0.97031(3.0%) 0.5119(+2.4%) 1.2878(14.1%) 2.63823(1.0%) 3.52551(2.5%) 0.97031(-3.0%) 0.5084(+1.7%) 1.1152(-25.7%) 2.57268(-3.5%) 3.38809(-6.3%) 4 29 cv cs x1% x0.1% 1 1/2 3/2 2.66518 3.61676 0.909381(-0.6%) 0.48511(-3.0%) 1.48752(-0.8%) 2.59586(-2.6%) 3.50864(-3.0%) 0.97775(-2.2%) 0.5095(+1.9%) 1.3440(-10.4%) 2.65240(-0.5%) 3.55990(-1.6%) 0.97775(-2.2%) 0.5115(+2.3%) 1.2061(-19.6%) 2.61539(-1.9%) 3.47005(-4.1%) 5 19 cv cs x1% x0.1% 1 3/4 3/2 3.49776 4.92515 0.98608(-1.4%) 0.72573(-3.2%) 1.47852(-1.4%) 3.56030(-3.9%) 4.71006(-4.4%) 0.95547(-4.5%) 0.7839(+4.5%) 1.1547(-23.0%) 3.44339(-1.6%) 4.74307(-3.7%) 0.95547(-4.5%) 0.7689(+2.5%) 1.0853(-29.4%) 3.3642(-3.8%) 4.57124(-7.2%) 6 29 cv cs x1% x0.1% 1 3/4 3/2 3.49776 4.92515 0.99072(-0.9%) 0.73903(-1.5%) 1.48737(-0.8%) 3.42364(-2.1%) 4.80974(-2.3%) 0.96663(-3.3%) 0.7724(+3.0%) 1.2649(-15.7%) 3.46916(-0.8%) 4.70729(-2.6%) 0.96663(-3.3%) 0.7580(+1.1%) 1.2038(-19.7%) 3.39449(-3.0%) 4.6