Functions to automatically determine and compute the relevant back-transformation for a rmult object.
backtransform(x, as=x)
backtransform.rmult(x, as=x)
gsi.orig(x,y=NULL)
gsi.getV(x,y=NULL)
For functions backtransform
or backtransform.rmult
, a corresponding matrix or vector containing the backtransformation of x
. Efforts are taken to keep any extra attributes (beyond, "dim", "dimnames" and "class") the argument "x" may have \
For function gsi.orig
, the original data with a compositional class, if it exists (or NULL otherwise). \
For function gsi.getV
, the transposed, inverse matrix of log-contrasts originally used to forward transform the original composition orig
to its coefficients/coordinates. If it does not exists, the output is NULL.
an rmult object to be backtransformed; for both gsi.*
functions: an rmult object to extract the relevant information from
an rmult object previously obtained with any compositional transformation of this package.
for both gsi.*
functions: an alternative object to extract the relevant information
from, in case that x
does not include it
R. Tolosana-Delgado, K.Gerald v.d. Boogaart http://www.stat.boogaart.de
The general idea of this package is to analyse the same data with different geometric concepts, in a fashion as similar as possible. For each of the four concepts there exists a family of transforms expressing the geometry in aan appropriate manner. Transformed data can be further analysed, and certain results may be back-transformed to the original scale. These functions take care of tracking, constructing and computing the inverse transformation, whichever was the original geometry and forward transformation used.
van den Boogaart, K.G. and R. Tolosana-Delgado (2008) "compositions": a unified R package to analyze Compositional Data, Computers & Geosciences, 34 (4), pages 320-338, tools:::Rd_expr_doi("10.1016/j.cageo.2006.11.017").
cdt
, idt
, clr
, cpt
, ilt
, iit
, ilr
, ipt
, alr
, apt
x <- acomp(1:5)
x
backtransform(ilr(x))
backtransform(clr(x))
backtransform(idt(x))
backtransform(cdt(x))
backtransform(alr(x))
Run the code above in your browser using DataLab