Alexander Kapeller
2011-May-20 21:25 UTC
[R] outout clarification of fitdist {fitdistrplus} output
Hello, I like to fit data against a negative binominal distribution x2<-c(rep(10,14),rep(9,8),rep(8,13),rep(7,11),rep(6,6),rep(5,18),rep(4,7),re p(3,21),rep(2,33),rep(1,55),rep(0,225)) f2<-fitdist(x2,"nbinom",method="mle") plot(f2) summary(f2) gofstat(f2) I receive the following result: Fitting of the distribution ' nbinom ' by maximum likelihood Parameters : estimate Std. Error size 0.3397148 0.03775577 mu 1.7787630 0.16429774 Loglikelihood: -701.0615 AIC: 1406.123 BIC: 1414.16 ------ I am wondering how to translate the size and the mu parameter into the usual number of successes (n) and the contiuous probability (p) . To me it is not clear what these mu and size parameter declares. Help is appreciated Best Alex [[alternative HTML version deleted]]
Ben Bolker
2011-May-21 15:50 UTC
[R] outout clarification of fitdist {fitdistrplus} output
Alexander Kapeller <mail <at> alexander-kapeller.de> writes:> > Hello, > > I like to fit data against a negative binominal distribution >[snip]> I am wondering how to translate the size and the mu parameter into the usual > number of successes (n) and the contiuous probability (p) . To me it is not > clear what these mu and size parameter declares. >See the help page for ?dnbinom, which gives the details of the different parameterizations ... Ben Bolker