[R] Fit a distribution to an "ECDF"

MyKenk michael.henk at gmail.com
Wed Nov 17 16:37:38 CET 2010


Actually, after thinking about this, what I want to do is to fit a
curve (Weibull and LogLogistic) to the first 9 data points in the
series.  The numbers represent percent of development, and I don't
"trust" the 1.000 point, so I want to fit these curves to get 2
estimates of the tail, both a thick and a thin.  I want to be able to
extrapolate percentage developed beyond 9 periods of development.

On Nov 16, 9:36 pm, Frank Harrell <f.harr... at vanderbilt.edu> wrote:
> Fitting curves to an ECDF will result in a fit that has the same precision as
> the ECDF if variances are calculated correctly.  So why not stop with the
> ECDF as your estimator?
>
> Frank
>
> -----
> Frank Harrell
> Department of Biostatistics, Vanderbilt University
> --
> View this message in context:http://r.789695.n4.nabble.com/Fit-a-distribution-to-an-ECDF-tp3045793...
> Sent from the R help mailing list archive at Nabble.com.
>
> ______________________________________________
> R-h... at r-project.org mailing listhttps://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guidehttp://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.



More information about the R-help mailing list