20/05/14 09:44:43.41 coRQ6//C.net
まあ、ベイズファクターの計算結果はさほど変わらんな。
> BEST::plotPost(js$r,compVal = 0,showCurve = F,col='lightgreen')
> lines(density(js$r),col='green')
> abline(h=0.5,col=4)
> fit.post=logspline(js$r)
> points(0,dlogspline(0,fit.post),pch=19,col='white',cex=1.5)
> points(0,dlogspline(0,fit.post),cex=1.5)
> points(0,dunif(0,-1,1),pch=19,cex=1.5)
> curve(dlogspline(x,fit.post),add=T)
> fit.F=logspline(atanh(r))
> curve(dlogspline(x,fit.F), col=2, add=T)
> d.prio=dunif(0,-1,1)
> d.post=dlogspline(0,fit.post)
> d.post/d.prio
[1] 2.917773
> dlogspline(atanh(0),fit.F)/d.prio
[1] 2.911407