[R-sig-Geo] normal score and back...

[G. Allegri]

I.ve approximated normal scores with the blom's method:
qnorm((rank(x)-0.375)/(sum(!is.na(x))+0.25))
but back-transormation using approx() doesn't work correctly, maybe
because of decimal values approximations.

Could you suggest me another way (then approx) to interpolate
(linearly) values to do back-t?

2008/2/26, G. Allegri <giohappy at gmail.com>:
> I'm sorry for asking again, but I haven't found an exhaustive answer yet.
>  Is there a package or a function in R to execute normal score
>  transformation (and back?).
>
>  I know GSLIB is an alternative, but I would prefare to make the whole
>  work inside R...
>
>  Thanks,
>  Giovanni
>