Seems for that approch M should me a matrix (which I bet it actually is in
this case)
not a data frame.
I wonder about the following, if you will forgive a slight digression:
It seems for really large matrices there could be a speedup as well as
savings of memory with something like.
M - (M0)
The short answer is that what you seem to want is the rq() default with
tau not specified. (Default is tau=.5).
In general rq() minimizes a sum of weighted absolute residuals. The
weights depend on tau (the conditional quantile of
interest), and turn out to be equal with tau = 0.5, i.e.,
One can always use the delta method (statistical differentials) for
problem of that sort, but in case the idea is to compute a confidence
interval, intervals based on asymptotic normality should be computed in
the scale where normality is most plausible, likely logit in this case,
and the
Nate,
I sent you some material off the list, before I saw your more detailed
discussion of quantile regression issue here on
the list. Here are a couple points for benefit of other participants.
1. Since the R quantile regression package supports nonlinear quantile
regression, so you
For combining imputation with PCA, transcan{Design} could be interest.
For imputation, I have been hearing about the mice program.
The documentation for princomp() suggests that a covariance matrix can be
entered via the covmat argument.
David F
r-sig-ecology-boun...@r-project.org wrote
In addition to the other comments, you may find it helpful to
look at using function derivative information in the nonlinear
regression section of Venables and Ripley.
David
[EMAIL PROTECTED] wrote on 10/29/2008 07:56:01 AM:
Dear R-ecology subscribers,
Here is a small contamination
This may be a long shot. You might try shutting down R and running your
script from a command (DOS) window using the RTERM.EXE utility.
Things do seem to run a little better that way as a rule.
David
ONKELINX, Thierry [EMAIL PROTECTED]
Sent by: [EMAIL PROTECTED]
10/10/2008 10:12 AM
To
Dulce,
Chapters dealing with R OOP and R/C interfacing can be found in
Chambers, J.M. 2008. Software for Data Analysis. Springer.
David
Dulce M. Bustamante [EMAIL PROTECTED]
Sent by: [EMAIL PROTECTED]
10/09/2008 04:47 PM
To
r-sig-ecology@r-project.org
cc
Subject
[R-sig-eco] Improving
Thanks for the illustration of xtabs.
A quibble: Doesn't the following work, substituting as.matrix() for
matrix()?
(Does seem to conserve the dimensions and dimension names.)
matrify-function(datatable, formula = units~site+spp, relativize=F){
tbl-xtabs(formula,data=datatable)
mx
