3

I'm looking to fit a bivariate Weibull distribution to my observations. I'm considering writing this myself but before starting I just want to check if anyone has come across a r-package including this functionality or similar.

I'm considering using the method proposed in this paper to estimate the parameters in the density function using maximum likelihood estimation: http://doi.org/10.18187/pjsor.v5i2.120

I'm already using the fitdistrplus package for univariate Weibull distribution fitting, but this only shows the marginal cases for a joint distribution. My observations include likely correlated data for two separate variables. The marginal distributions follow Weibull shapes though, which is why I feel confident that a bivariate Weibull distribution would be applicable.

Johannes M
  • 31
  • 4
  • I guess you already found http://www.real-statistics.com/distribution-fitting/fitting-weibull-regression/ - there are quite some traps when you try to do it yourself, believe me. I tried doing that years ago... – B--rian May 27 '20 at 08:17
  • I’m voting to close this question because this might already have an answer at https://stats.stackexchange.com/questions/346249/fitting-weibull-distribution-in-r – B--rian May 27 '20 at 08:18
  • On 2nd thought, the linked stats-post might not help you since it deals with univariate Weibull distribution. Nevertheless, you might re-post the question on that site. – B--rian May 27 '20 at 08:22
  • 1
    Thanks, B--rian! It's fairly easy to find ways in R to fit univariate Weibull distributions, but I haven't found any for multivariate. Already tried to post in CrossValidated, but they said it was off topic to ask for tips on r-packages there. – Johannes M May 27 '20 at 08:33

0 Answers0