Life Data Analysis

#1
Good Morning,

This is the first time that I have posted anything and I am really in need of guidance.

I am trying to figure out the best way to predict failure using warranty data. I have a product that had very few claims in the beginning, but the longer the life of the product the more claims we received and it turned out to be a huge issue.

Is there anyway to predict failure when data doesn't fit the weibull distribution and we have only a few claims to make the analysis?

Any assistance would be helpful! I am completely lost on how to solve this problem.

Thank you!
 

Link

Ninja say what!?!
#2
You have to explain your data in more detail. What kind of variables do you have? Do you only have claims from failed devices? Do you have the time of purchase, allowing you to compute it's "survival" time? There are plenty of ways to approach survival analysis without using a weibull dist.
 
#3
The data comes out in an excel format with the following information:

settlement month - when we paid for the claim
# of models produced
date it was produced
# of models purchased
months in service - when it failed

All this information is warranty information

I would love to learn ways to approach this problem.

Thank you!
 
#4
The data comes out in an excel format with the following information:

settlement month - when we paid for the claim
# of models produced
date it was produced
# of models purchased
months in service - when it failed

All this information is warranty information

I would love to learn ways to approach this problem.

Thank you