[R] MLE for censored distributions in R

[Vincent Goulet]

Le mar. 22 janv. à 17:47, Thomas Downey a écrit :

>
> Hi just wondering if there is a package that can get the maximum  
> likelihood
> or method of moments estimator for distributions with censored  
> data?  The
> distributions I'm interested in are: Exponential, pareto, beta,  
> gamma and
> lognormal.

I'm highly biased here, but you may have a look at package actuar. The  
function coverage() will define the pdf or cdf of a random variable  
modified in many sorts of ways (truncation, censoring, inflation,  
etc.) based on any distribution you want. You can then use the  
function is usual fitting procedures such as fitdistr() for ML. See  
the "lossdist" vignette in the package for details.

HTH

---
   Vincent Goulet, Associate Professor
   École d'actuariat
   Université Laval, Québec
   Vincent.Goulet at act.ulaval.ca   http://vgoulet.act.ulaval.ca