WOLFRAM

Wolfram 语言与系统参考资料中心

比较非参数式和参数式可靠性模型

分别使用 SurvivalDistribution 和 CensoredDistribution 比较生产出来的绳索的可靠性函数的 NPMLE 和 MLE 估计.

    
Wolfram Language code: data = {100, 43, 100, 100, 63, 26, 100, 100, 100, 100, 100, 100, 100, 69, 58, 99, 95};
Wolfram Language code: pars = FindDistributionParameters[data, CensoredDistribution[{0, 100}, WeibullDistribution[a, b]]]; sf[x_] = SurvivalFunction[CensoredDistribution[{-Infinity, 100}, WeibullDistribution[a, b]], x] /. pars;
Wolfram Language code: 𝒟 = SurvivalDistribution[EventData[data, Boole[# ≥ 100]& /@ data]];
Wolfram Language code: Plot[{SurvivalFunction[𝒟, x], sf[x]}, {x, 0, 120}, PlotPoints -> 100, Exclusions -> None, Filling -> {1 -> {2}}, PlotStyle -> Thick, Frame -> {{True, False}, {True, False}}, FrameLabel -> (Style[#, FontFamily -> "Verdana", Bold, 12]&) /@ {"Force", "Reliability"}, ImageSize -> 500]