relation of data points to distribution using Competing Risk Mixture option of Life Distribution platform

Question

Hi All,

I use Life Distribution platform from Reliability and Survival option to fit distribution to reliability results.

To distinguish between failure causes I use Competing Risk Mixture option in Life Distribution platform (see graphs below and attached file).

JMP provide parameters of fitted distributions but I have no clue what data points have been used for each fitted distribution.

Is there way to find out what data point have been used to fit each distribution?

Thanks.

data.jmp

ptolomey · Answer

Hi,

Thank you for response.

You right, my question was: "which data points is the Loglogistic fitted to and which data points is the Frechet fitted to"?

Can I use following relate each data point  to distribution using following expression:  Min {   [  Actual_Y(X)  -  Loglogistic_Y(X)  ] ,    [  Actual_Y(X)  -  Frechet_Y(X)  ]   }

Thanks.

Phil_Kay · Answer

Hi,

I think that you are asking: "which data points is the Loglogistic fitted to and which data points is the Frechet fitted to?"

My understanding is that all data points are used to fit the aggregate of the Loglogistic and Frechet distribution.

That is, each observation arises from the processes described by the Loglogistic distribution and the Frechet distribution.

So you can't really separate the data points in the way that you are asking.

Having said that, you can see the contribution from each process in the plot on the right. E.g. before 20 hours the process described by the Loglogistic distribution has no contribution. That is, if the process described the Frechet distribution were eliminated, you would expect no failures before 20 hours.

Does that make sense? Does that help?

Regards,

Phil