Hello,
I'm having an issue with some lifetime data I have in which I am trying to generate a lifetime extrapolation using the Box-Cox transformation within the Fit Life by X platform. My end goal is to plot time-to-fail data from 3 different manufacturing lots that have all been tested to failure at varying voltages. I need the ability at each stress voltage to get the μ and σ from each manufacturing lot (within lot variation) and then to average the 3 values to return an average across the 3 lots for the purpose of my lifetime modelling.
Currently my issue exists whereby JMP groups all datapoints from the 3x lots at a particular voltage into one combined dataset, thus giving me a much larger σ than if I was to get the within lot σ of each lot and then average that across the 3. I am setting up my analysis with the following columns (also available in the attached .jmp file):
Y, Time to Event: TTF (Hours)
X: Stress Voltage
Censor: Censor
From what I can see, this method groups all datapoints at a particular voltage into a distribution, even though the data is from 3x different lots. If I add Material Type to the By column, it will split out the σ statistic per lot but I'm unsure how to then take the 3 different σ values, average them and then apply the averaged value to my lifetime extrapolation. I have colleagues who are working with R and Matlab and are able to do the above analysis request quite easily so I'm wondering if there is just a simple change I can make in JMP to imitate this analysis. Any help would be appreciated.
