864024_en-US

liliconcom

Life distribution: how to set the markers and weibull fit to use Kaplan-Meier or Turnbull estimates

Hello,

 

The below is the test table I made.

According to the description: https://www.jmp.com/support/help/en/18.1/index.shtml#page/jmp/event-plot.shtml#ww1386833

It meas there are total 100 samples.

And exactly at 4 readings: 50, 100, 150 and 200, there are 25 new failures happened.

So the failure rate I expected will be:

50, 25%

100, 50%,

150, 75%

200,100%.

 

 

The markers in the Weibull plot are very different from the failure rate I expected.

For example, the first read point at time 50 has failure rate as ~13%.

Now I learned that this is because the life distribution is using the Midpoint estimate to do the plot and fit.

But I want to use the Kaplan-Meier estimate to plot the markers and fit weibull, which is more intuitive to my audience.

 

Is there a way to change that in JMP? I had no luck searching in the preference setting.

 

Thank you very much.

I am using JMP18.2 now.

Basic Data Analysis and ModelingData Exploration and VisualizationReliability AnalysisWindowsRe: Life distribution: how to set the markers and weibull fit to use Kaplan-Meier or Turnbull estimates

Looks like a follow up to Life distribution: The failure rate/probability plot does not match with the table 

 
Please have a look at Analyze/Survival:

 

 

As an alternative, you can use GraphBuilder. In the Summary Statistics menu you can find "Cumulative Percent" - which uses the Kaplan-Meier definition to calculate the probability. With a trick *), you can even use it for points [fromAdd CDFs to Graph Builder]:

Please note: 
for i/N, the last points has a probability of 1 ... and gets lost when probability scales are used on the Y axis - like in the plots in this post. This is why I submitted a wish to add alternative probability functions as an option in Graph Builder:Summary Statistics in Graph Builder - add Cumulative Probability 

*)

Graph Builder(
	Transform Column( "1", Formula( 1 ) ),
	Size( 429, 272 ),
	Show Control Panel( 0 ),
	Summary Statistic( "Median" ),
	Graph Spacing( 4 ),
	Variables( X( :Start time ), Y( :"1"n ) ),
	Elements(
		Points( X, Y,  Summary Statistic( "Cumulative Percent" ) ),
		Line(
			X,
			Y,
			Connection( "Step" ),
			Summary Statistic( "Cumulative Percent" )
		)
	),
);

Re: Life distribution: how to set the markers and weibull fit to use Kaplan-Meier or Turnbull estimates

Thank you Hogi.

 

I did not find a way to direcly use JMP life distribution to plot and weibull fit using Kaplan-Meier estimate.

I will have to go back to Matlab for the plot and weibull fit.