Hi,
I?ve came across something that I can?t explain and I would appreciate if
anyone could have a go at it.
In the library ?fitdistplus? there is a function ?descdist? to help on the
decision of choosing a distribution to fit. The same function also allows
bootstrap this is to take in account the uncertainty of the calculated
values.
If you run the line below a few times you will find quite a big variation
each time you run it, in fact that?s what is made me suspicious. If I?m
bootstrapping always from the same distribution curve why do I have such
variation?
A second question that arises to my mind and this probably is due to the
lack knowledge on the subject. But if the Cullen and Frey graph is to help
on the decision on which distribution to choose is the line below just
giving me the uncertainty of the estimates of Kurtosis and Skewness and
should I ignore all the lines in the graph as I already fitted a weibull
distribution to the original data?
library("fitdistrplus")
descdist(rweibull(1000,shape=13.74286,scale=38.07489),boot=1000)
Cheers
Patrao
--
View this message in context:
http://r.789695.n4.nabble.com/cullen-and-Frey-graph-in-fitdistrplus-tp3924732p3924732.html
Sent from the R help mailing list archive at Nabble.com.