Arvind
April 10th, 2007, 08:00 PM
Hi ,
In continuation to my previous post "How to predict warranty returns" i managed to do warranty analysis by treating units of each build month as a seperate/unique population.
Assuming that i subject all build units of a month to field test (say Dec 05 build month units) & then i use times to failure with suspensions format(based on successive month returns ) & predict the probability of failure at the end of 18 months (warranty period).
Is this method correct (sometimes i get erratic values) ??????
Can anyone tell me what is the difference between "Failure rate" & "probability of failure" in the QCP. If i want to estimate the total failures in 18 months for these build month units, which one do i need to use.
Please Ref to the attached file.
Can any one suggest me in this regard.
Thanks
Arvind.
In continuation to my previous post "How to predict warranty returns" i managed to do warranty analysis by treating units of each build month as a seperate/unique population.
Assuming that i subject all build units of a month to field test (say Dec 05 build month units) & then i use times to failure with suspensions format(based on successive month returns ) & predict the probability of failure at the end of 18 months (warranty period).
Is this method correct (sometimes i get erratic values) ??????
Can anyone tell me what is the difference between "Failure rate" & "probability of failure" in the QCP. If i want to estimate the total failures in 18 months for these build month units, which one do i need to use.
Please Ref to the attached file.
Can any one suggest me in this regard.
Thanks
Arvind.