Hello, I also try to fit a skewed distribution (like skewed student t) to data points. Do you have an idee howto do this??? thank you fabrice dusonchet *********************************************************************************** This email and any files transmitted with it are confidentia...{{dropped}}
>>>>> "Fabrice" == Fabrice Dusonchet <fdusonchet at eim.ch> >>>>> on Wed, 4 Aug 2004 18:55:18 +0200 writes:Fabrice> I also try to fit a skewed distribution (like Fabrice> skewed student t) to data points. Do you have an Fabrice> idee howto do this??? One recommended approach would be to use the mle() functionality in the standard package 'stats4' (was package 'mle' in 1.8.x). It uses optim() but provides several extra niceties. Bonnes salutations de Zurich! Martin Maechler, Seminar f??r Statistik, ETH Zurich
>I also try to fit a skewed distribution (like skewed student t) to data >points. Do you have an idee howto do this???library(skewt) y <- rskt(500,2,2) # simulate 500 observations from the skew t distribution with df=2 and gamma=2 skewtmle <- function(df,gamma){ return(-sum(log(dskt(y, df, gamma)))) } fit <- mle(skewtmle,start=list(df=1,gamma=3),method="L-BFGS-B",lower=c(1e-8,-Inf)) summary(fit) logLik(fit) vcov(fit) plot(profile(fit), absVal=F) confint(fit) Hannu Kahra Progetti Speciali Monte Paschi Asset Management SGR S.p.A. Via San Vittore, 37 - 20123 Milano, Italia