请问如何计算某个分布的X0.00135分位数?

[#1]

当计算过程能力时, 数据不服从正态分布时. 需要找出服从的其它分布(比如 weibull分布)的X0.00135分位数.

请问在JMP里, 如何计算出如下图这个分布的X0.00135分位数? 

[#2 · Accepted solution]

Hi @Xinghua ,

  There are a couple different ways you can do this.

  From the Distribution platform, you can click the red hot-button next to the name of whatever you're analyzing (Purity), select Display Options and Custom Quantiles.

This will give you a popup window that allows you to enter whatever custom quantiles you want.

  Or, if you have JMP v19, once you've fitted the data to get your continuous distribution, you can click the red hot-button next to Fitted Weibull Distribution, and select Fitted Quantiles, which gives you a similar popup window.

Hope this helps!,

DS

[#3]

Thank you very much.

[#4]

我刚刚尝试了一下, JMP给出的分位数以及自定义的分位数, 是对于数据的(从小到大排列), 而不是对于数据所服从的分布. 我需要的是分布的分位数(以分布的面积计算).

如下图, X=99.73左侧红色区域的面积占分布的5%.

[#5]

Hi @Xinghua : As I indicated in my previous post, you can use the quantile profiler under the Fitted Weibull Distribution output if you want the quantile based on the fitted Weibull distribution.

Edit: Just click on the red triangle next to "Fitted Weibull Distribution" and choose Profilers/Quantile Profiler. Then in the Quantile Profiler, input 0.00135 on the X-axis labeled Probability. The quantile will then be displayed on the Y-axis.

[#6 · Accepted solution]

Hi @Xinghua : You can also use the quantile profiler under the Fitted Weibull Distribution output if you want the quantile based on the fitted Weibull distribution.

[#7]

你可能真正的解决了我的问题 --- 寻找分布的某个分位数 (而不是数据的分位数). 非常感谢!