3
$\begingroup$

I am doing a project in which I need to estimate Weibull parameters for car part failures (I know the data follow Weibull). I have data for 1000 cars (part failure data). Now the problem is suppose some part fails after 6 years (from date of manufacturing) . But I have data for only the first three years. Now in three years this particular part may have failed in only a very few cars. Since I am taking miles driven as cycle to estimate parameters, very few will be less than some threshold. So when I estimate parameters based on that sample, it gives me wrong results. Can anyone help me how to do this?

$\endgroup$
1
  • 1
    $\begingroup$ It sounds like a good part of your data is right-censored. Look up survival analysis and right-censoring. A maximum likelihood approach might work for you here. $\endgroup$
    – soakley
    Dec 19, 2013 at 14:26

1 Answer 1

1
$\begingroup$

@soakley already gave you the main keyword - survival analysis - so I'll just mention that a good solution is either R's classic survival package (this function), or flexsurv, which has more flexibility, but a different parametrization for Weibull (flexsurvreg function here).

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.