Meng,
This really comes down to what question you are trying to answer. Before
worrying about details of default contrasts and issues like that you first
need to work out what is really the question of interest. The main
difference between declaring a variable ordered or not is the default
Here are a couple of approaches:
dftest-data.frame(x=1:12, y=(1:12)%%4, z=(1:12)%%2)
x_test=c(x,y)
aggregate( dftest[,x_test], dftest['z'], FUN=mean )
z x y
1 0 7 1
2 1 6 2
### Or
tmp.f - as.formula( paste( 'cbind(',
+ paste( x_test, collapse=',' ),
+ ') ~ z' ) )
aggregate( tmp.f,
The pairs2 function in the TeachingDemos package does what you describe.
You give it 2 matricies instead of just one and it creates the plots.
On Wed, May 8, 2013 at 10:49 AM, Adel adelda...@gmail.com wrote:
Dear list-members,
I wonder if there is a way of creating a scatter plot table/grid
In addition to David's great answer also read
?contr.SAS
?contr.treatment
?relevel
and also sections 4 and 11.1.1 in An Introduction to R.
On Fri, May 3, 2013 at 9:47 PM, David Winsemius dwinsem...@comcast.netwrote:
On May 3, 2013, at 3:32 PM, Iuri Gavronski wrote:
Hi,
I would like to
The offset(x) term means to include x (or whatever is in the parentheses)
into the model as is, without computing a slope for that term. You could
also include offset( 1 * x ) instead which might make this a bit more
explicit (but would not actually make any difference). Since x by itself
is
This is FAQ 7.21
The most important part of that answer is the last few lines where it says
in effect Don't Do This and shows the basics of using a list instead.
On Mon, Apr 29, 2013 at 5:07 AM, Rui Esteves ruimax...@gmail.com wrote:
Hello,
This is very basic and very frustrating.
Suppose
Jim Lemon showed one option of keeping a list of the files that have
already been processed and comparing to this list to see which files are
new. Here are a couple of other possibilities:
You could move a file from the input folder to an archive folder after
processing it (file.rename or system
This link
http://math.stackexchange.com/questions/18488/diagonal-of-an-inverse-of-a-sparse-matrix
might
help. It is about sparse matrices, but the general idea should be able to
be extended to non-sparse matrices as well.
On Fri, Apr 19, 2013 at 8:13 AM, Camarda, Carlo Giovanni
To further David's comment, just think what the world would be like if
Alexander
Fleming had discarded an obvious outlier in 1928 (
http://en.wikipedia.org/wiki/Penicillin#Discovery).
On Tue, Apr 9, 2013 at 7:46 AM, David Winsemius dwinsem...@comcast.netwrote:
On Apr 9, 2013, at 4:12 AM,
On Mon, Apr 1, 2013 at 10:26 AM, R. Michael Weylandt
michael.weyla...@gmail.com wrote:
Yes -- in fact that's the way you're supposed to do it and the
RStudio way you mentioned above is a shortcut.
And the longest distance between any 2 points is a shortcut.
--
Gregory (Greg) L. Snow Ph.D.
There is the pairs2 function in the TeachingDemos package that works like
cor(x,y) where you give it 2 matricies/data frames and it gives the
pairwise plots between the 2 groups. There is currently not a formula
interface.
On Fri, Mar 29, 2013 at 12:16 PM, Michael Friendly
The tkexamp function in the TeachingDemos package has an animation control
that you can use.
The animation package will create video files of animated graphs.
On Wed, Mar 27, 2013 at 7:20 AM, catalin roibu catalinro...@gmail.comwrote:
Hello all!
I want to create animated chart of
You do exactly the same thing, just in the box in the lower left corner
(just above the Apply button) you need to scroll down and select editorbg
for the background and editortext for the text color.
On Fri, Mar 22, 2013 at 8:14 PM, Pfeiffer, Steven (pfeiffss)
pfeif...@mail.uc.edu wrote:
The TeachingDemos package has %% and %=% functions that can be chained
simply, so you could do something like:
sum( 5:1 %=% 1:5 %=% 10:14 )
and other similar approaches.
The idea is that you can do comparisons as:
lower %% x %% upper
instead of
lower x x upper
On Mon, Mar 18, 2013 at
The my.symbols and ms.arrows functions in the TeachingDemos package may
help.
On Mon, Mar 18, 2013 at 3:50 AM, Jie Tang totang...@gmail.com wrote:
hi R users:
I have a dataset including u wind in x-axis and v wind in y-axis.
How can I plot the u,v wind data in vector or barb figure?
I think that the better approach in R, if you really need to do this, is to
convert your array to a list, assign the names to the list, then attach the
list to the search path so you can access the variables by name. For
example:
x - c(1,2,3)
x2 - as.list(x)
names(x2) - c('a','b','c')
I don't see any mention of odfWeave yet. It works with OpenOffice files.
OpenOffice is a free equivalent to MS Office and can read and write Word
documents. So you can create your template file in OpenOffice (or use MS
word and convert to OpenOffice format, OO will read word docs, and recent
You can use the lapply or rapply functions on the resulting list to break
each piece into a list itself, then apply the lapply or rapply function to
those resulting lists, ...
On Mon, Mar 11, 2013 at 3:41 PM, Not To Miss not.to.m...@gmail.com wrote:
Thanks. That's just an simple example - what
The only real improveent I can see over Ivan's solution is to use lapply
instead of the loop (this may just be person preference though).
Something like:
list_df - lapply( lista_rea_c, function(x) read.xls( file=
paste0(path,x,/,x,.xls),1,header=TRUE,as.data.frame=TRUE))
my_df - do.call(rbind,
In addition to the other suggestions that you have received you may want to
look at the spread.labs function in the TeachingDemos package and the
spread.labels function in the plotrix package. These only spread in 1
dimension, but may be able to do what you want. The thigmaphobe.labels
function
Instead of read the manual, how about read the answers that you have
already received. Or tell us why those answers are not good enough. Or
you can read the manual on the .filled.contour function (which is on the
same page as filled.contour).
On Fri, Mar 8, 2013 at 10:05 AM, Jing Lu
For anyone looking for an intermediate to advanced level short course on R
this summer, I will be presenting at the BYU Summer Institute of Applied
Statistics June 19-21, 2013.
The official page is here:
http://statistics.byu.edu/r-beyond-basics-38th-annual-summer-institute-applied-statistics
Try the following to see if it does what you want:
## init
con - textConnection(output, w)
options(echo=FALSE)
sink(con)
addTaskCallback( function(expr, out, err, vis) {
sink()
close(con)
if(vis) {
cat(paste(#, output, collapse=\n), \n)
}
con - textConnection(output, w)
sink(con)
TRUE
})
3+4
If you want to receive all the e-mails, but don't want your e-mail program
to pop up a message every time an e-mail from r-help comes in then you can
configure your e-mail client to automatically move all the e-mails from
r-help to a specific folder/label/etc. and that should make it so you can
The HWidentify and HTKidentify functions in the TeachingDemos package let
you specify a label for each point, then that label is displayed when you
hover over that point with the mouse (and it goes away when you move the
mouse away from that point).
On Thu, Feb 28, 2013 at 8:36 AM, Agustin Lobo
Nicole,
Since you seem more interested in accusing David of being rude than
recognizing your own rudeness and taking steps to overcome that and
increase your chance of getting useful responses I will quote a few lines
from the posting guide for you (the entire posting guide is available from
the
What have you tried so far?
The points function will add points to an existing graph (base) and the
text function will add text to an existing graph.
If those don't do what you need then give us some more details.
On Tue, Feb 26, 2013 at 2:04 PM, Debs Majumdar debs_st...@yahoo.com wrote:
Hi
Have you queried the value of 'cex' and related parameters at the different
time points?
The help page for par says that when you set mfcol or mfrow that cex is
changed, but I don't know if the layout function also changes those or not.
I would start by peppering your code with calls to
Or if David's answer seems like too much work you could use the `mvrnorm`
function in the MASS package to generate 2 vectors with the given
correlation and sample size and feed those vectors to the `cor.test`
function.
Or Pearson's test can be computed in 1 line of R code without needing any
Part of the space between the boxplots is the margin area, you can remove
that. You can also adjust the width of the box within the plot:
layout(matrix(c(1,2,3), 3, 1, byrow = TRUE), heights=c(0.3,0.3,0.6))
op - par(mar=c(0,4,0,2)+0.1)
boxplot(rnorm(100), horizontal=TRUE, axes=FALSE, width=1)
will give it a try.
** **
Petr
** **
*From:* Greg Snow [mailto:538...@gmail.com]
*Sent:* Friday, February 22, 2013 6:45 AM
*To:* PIKAL Petr
*Cc:* r-help
*Subject:* Re: [R] package ReadImages
** **
Some possibilities: The EBImage package on Bioconductor; the jpeg and png
, 2013 at 6:13 PM, Greg Snow 538...@gmail.com wrote:
To know for sure we need to know how you are running these different R
sessions, but here are some possibilities:
The help page for set.seed says that if no seed exists then the seed is
set based on the current time (and since 2.14.0
Some possibilities: The EBImage package on Bioconductor; the jpeg and png
packages read jpeg and png images.
On Thu, Feb 21, 2013 at 8:26 AM, PIKAL Petr petr.pi...@precheza.cz wrote:
Dear all
I prepared some image processing routine which depended on package
ReadImages. Basically I
To know for sure we need to know how you are running these different R
sessions, but here are some possibilities:
The help page for set.seed says that if no seed exists then the seed is
set based on the current time (and since 2.14.0 the process ID). So one
possibility is that 2 of the sessions
Have you plotted the data and the lines to see how they compare? (see
fortune(193)).
Is there error around the line in the data? The nls function is known to
not work well when there is no error around the line. Also check and make
sure that the 2 methods are fitting the same model.
You
The abline function works fine for simple linear regression because there
is only 1 line, but with multiple linear regression there are an infinite
number of lines and you need to decide which to plot (or find a way to plot
t he plane/hyperplane/surface/etc.).
One option is to use the
?sprintf
On Thu, Feb 14, 2013 at 6:59 AM, Esam Tolba eato...@gmail.com wrote:
Dear all
I am using r (2.15.2) under windows 7 32bit.
I want to use system command to execute an external program inside a
for loop. this program needs the input and output names.
since the names will change in
Look at the gsubfn package, it has functionality to do this.
Here is an example from the vignette:
fn$cat(pi = $pi, exp = `exp(1)`\n)
pi = 3.14159265358979, exp = 2.71828182845905
The formula syntax to function conversion in the package may help with the
last part.
On Fri, Feb 8, 2013 at
If you are on windows then there is a task scheduler for windows.
Generally the best solutions are to use the OS tools to schedule your task,
chron or task scheduler.
If you need a solution inside of R then you can use the tclTaskSchedule
function in the tcltk2 package.
On Mon, Feb 11, 2013 at
The thread starting with this post:
https://stat.ethz.ch/pipermail/r-sig-teaching/2013q1/000534.html answers a
similar question and many of the answers there should apply here as well.
On Tue, Feb 12, 2013 at 11:51 AM, David Arnold dwarnol...@suddenlink.netwrote:
All,
If you have any good
See ?commandArgs for a way to access the arguments on the command line.
With that you can write the FunkyCold.R script to grab the command line
arguments and run the proper function (after loading your package). If you
give more detail on what you want to accomplish we may be able to give more
The stats package has the optimize function that could be used for this.
On Tue, Feb 5, 2013 at 2:08 PM, John Sorkin jsor...@grecc.umaryland.eduwrote:
I am looking for a package that will allow me to choose R (a real number)
that minimizes the correlation of y and x^R, i.e.
find R such that
An often better approach is to use a function like with or within. These
allow you to run a command with your list (data frame, environment, etc.)
in the first position of the search list so they behave like they are in
the global environment (actually better because they temporarily mask
rather
It is strongly discouraged in R to have functions that change data values
in the global workspace (or any location other than their local
environment).
The usual procedure in R is to have your function return a modified version
of the object and the user then decides what to do with it. They can
Ivan,
In reference to your part 2), in 1989 Li and Duan published a paper where
they examined the effect of using the wrong link function (
http://projecteuclid.org/DPubS?service=UIversion=1.0verb=Displayhandle=euclid.aos/1176347254).
The short version is that they found that in common models
You could always just run something like
system(path/to/R/RScript myfile.R)
from within R.
But this also sounds like something that the parallel package may be
helpful with to use your 8 cores to speed up your update process.
On Thu, Jan 31, 2013 at 2:12 PM, Brigid Mooney bkmoo...@gmail.com
This is FAQ 7.21. The most important part of that answer is at the end
where it says that it is better to use a list. Your code could be
something like:
plotroc - list()
for (i in levels(mergeTrn$Continent) {
# matrix defined here
plotroc[[ paste(plotroc_GBM_TRN_,i, sep=) ]] - matrix
}
now
If you want to convert between different units using base graphics then
look at the grconvertX and grconvertY functions (in the graphics package).
These functions will convert from/to user coordinates, inches, device,
figure, and plot coordinates. So you could use grconvertX to find out
what
I don't see a symbols function in the gtools package, do you mean the
symbols function in the graphics package?
If so, there is not a simple legend or key function to create the legend
(the number of possible options would make it more complicated than
building the legend by hand). You will need
You could use the strapply function from the gsubfn package to extract the
data from strings. This will return a list that you could use with
do.call(rbind(
The stringr package may have something similar or an alternative (but I am
less familiar with that package).
On Thu, Jan 17, 2013 at 9:21
The important thing to understand is that $ is a shortcut for [[ and you
are moving into the realm where a shortcut is the longest distance between
2 points (see fortune(312)).
So your code can be something like:
state - 'oldstate'
balance - 'oldbalance'
dataa[[balance]][ dataa[[state]]=='AR' ]
The tables package may be of use to you for this.
On Fri, Jan 11, 2013 at 4:17 AM, Pancho Mulongeni
p.mulong...@namibia.pharmaccess.org wrote:
Hi, I have a dataframe with n columns, but I am only looking at five of
them. And lots of rows, over 700.
So I would like to find frequencies for
I don't know of any tools within R to do something like this, but one
possibility is to use other software that will interface with your device
and can then talk with R, one possibility for an intermediate program is
here: http://www.windmill.co.uk/index.html.
On Fri, Jan 11, 2013 at 4:41 PM,
There is also the my.symbols and ms.polygon functions in the TeachingDemos
package that have syntax similar to the symbols function:
library(TeachingDemos)
x - runif(10)
y - rnorm(10)
my.symbols(x, y, ms.polygon, inches=0.3, n=400, add=FALSE, lty=3)
On Sat, Jan 12, 2013 at 7:42 PM, Ved P.
This is fairly straight forward, you can do something like:
myfiles - scan('files.txt', what=)
base - sub(\\.[^.]*$, , myfiles)
lapply( seq_along(myfiles), function(i) {
mydat - read.table( myfiles[i] )
sink( sprintf('%sSummary.txt', base[i]) )
summary(mydat)
sink()
bmp(
If you have a working function that gives the power for a given sample size
then you can use the uniroot function to find which sample size will give
the desired power (many of the standard power functions use this internally
when computing anything other than power). You just create a function
You might be better off using the knitr package which has options for
whether to evaluate code or not (or which pieces of code to evaluate). If
you use this in Rstudio then you can knit directly to html from a variety
of input formats. Otherwise you can use the external program pandoc to
convert
The apply function takes another function as the argument to be applied to
each row (or column) of the matrix. You can use an existing function like
you did with 'max' and in this case there is an existing function called
'which.max' that does exactly what you want. If there is not an existing
To further the understanding of the loess fit and how the tricube weight
work you may want to look at the loess.demo function in the TeachingDemos
package. It will create a scatterplot of the data and show the loess fit,
then when you click on the plot it will show the weights used for
predicting
I believe the problem could be that xyplot uses grid graphics and plot.new
and curve are base graphics functions and the 2 graphics systems (grid and
base) don't play nicely together without a little extra work. In general
the gridBase package helps them play nicely, but I am not sure that it
Not as a simple color, but you can use the rasterImage function to add a
rectangle with a checkerboard pattern to a an existing plot, the examples
for the function use a simple checkerboard pattern.
Be careful of overusing something like this, too many patterned areas in a
plot can be more
And look at the grconvertX and grconvertY functions for possibly ways to
compute positioning information when using text with base graphics.
On Mon, Jan 7, 2013 at 11:35 AM, David Winsemius dwinsem...@comcast.netwrote:
On Jan 7, 2013, at 8:06 AM, Michael Rennie wrote:
Any thoughts on what
In current versions of R the apply functions do not gain much (if any) in
speed over a well written for loop (the for loops are much more efficient
than they used to be).
Using global variables could actually slow things down a little for what
you are doing, if you use `-` then it has to search
If you want help from us mere mortals then it would help us to help you if
you could tell us not only the number of commands that you tried but what
the actual commands were.
Also the phrase none worked does not help us help you, see fortune(324).
A description of what happened and how the
The only way that I can think of that would not use iteration would be to
create some type of lookup table with every possible vector you may ever
use, then lookup the results in that table.
If you are happy with internal iteration then one possibility is:
x - c(4,3,5)
There is also another fortune:
fortune('toad')
The problem here is that the $ notation is a magical shortcut and like any
other magic if used incorrectly is likely to do the programmatic equivalent
of
turning yourself into a toad.
-- Greg Snow (in response to a user that wanted to access
Others have mentioned the readability issues (which are important), but
think about what would happen in the case where the variable lambda rule
ended up with a value like:
lambda.rule - a;system('SomeEvilSytemCommand')
Or what if the element of the list desired has a space in its name.
And
The idiom of 'par(new=TRUE)' should be hidden, it is never the first thing
that you should try and is only rarely needed and as you have seen often
causes more problems than it helps.
First you should ask yourself if using 2 different coordinate systems in
the same plot is the best approach, you
It would help if we had some idea what data_min1$macdhist actually is. I
would not expect to see that error if it is a simple numeric vector, so it
probably has a class of some sort which causes the plot command to call a
specific method (but we don't know which one) which may not have a col
Using pch you can use all the symbols in the current font, try:
plot(0:15, 0:15, type='n')
points( (0:255)%%16, (0:255)%/%16, pch=0:255 )
then do it again with
points( (0:255)%%16, (0:255)%/%16, pch=0:255, font=5 )
(font 5 is usually a symbol font, fonts 2, 3, and 4 are bold and italic
versions
(0,15000))
** **
With a simple test like f(x)=x*x the ylim-option works fine.
So I guess, that it will be something with kind of data set.
** **
A Pro-answer would be great.
** **
Thx a lot
** **
Tom
** **
*Von:* Greg Snow [mailto:538...@gmail.com
Look at the tapply function. Possibly using round or cut if needed. Other
tools that may be useful (but probably overkill for the simple problem)
include aggregate, ave, and the plyr package.
On Sat, Dec 15, 2012 at 9:12 PM, Elaine Kuo elaine.kuo...@gmail.com wrote:
Hello
I have a table
One option is to put all the data frames into a list then use the Reduce
function.
On Mon, Dec 17, 2012 at 9:03 AM, Vasilchenko Aleksander
vasilchenko@gmail.com wrote:
Hello.
I have for example 4 or more dataframe which like such this example:
date value
2006-11 0.4577647
try ylim=c(0,1500)
You are missing the c which constructs a vector, without the c it gets
confused.
On Sat, Dec 15, 2012 at 5:14 PM, Tom Hoffrichter
tom.hoffrich...@googlemail.com wrote:
Hi everybody,
just arrived at R and immediately I got a problem.
Here's my script:
Another option (better or worse probably depends on many things) is to use:
b - head(a, -1)
On Wed, Dec 12, 2012 at 1:14 AM, e-letter inp...@gmail.com wrote:
On 12/12/2012, Daniel Nordlund djnordl...@frontier.com wrote:
-Original Message-
From: r-help-boun...@r-project.org
It may be overkill, but you can specify the model pieces using the offset
function in the model, then the predictions work out (at least for my
simple trial case). Something like:
fit - glm( y ~ 0+offset(-1 + 2*x), family=binomial, data=data.frame(y=0,
x=0) )
predict( fit,
I think the error is because the other functions are expecting the plot.new
function to do some specific things to set up the graphics device for a new
plot, but your function did not do those things.
You might consider the trace function as an alternative to what you are
trying. It can be used
?findInterval
On Thu, Dec 6, 2012 at 3:00 PM, Jonathan Greenberg j...@illinois.eduwrote:
Folks:
Say I have a set of histogram breaks:
breaks=c(1:10,15)
# With bin ids:
bin_ids=1:(length(breaks)-1)
# and some data (note that some of it falls outside the breaks:
Read the ?Startup page, it documents the files that are read on startup and
how to skip them, it may suggest another way to start a fresh session
without deleting files to see if that works for you, it also gives other
files or sources that you may investigate to find the differences in the 2
look at functions replicate and mvrnorm functions (the later in the MASS
package).
On Sat, Dec 1, 2012 at 12:02 PM, mboricgs mbori...@hotmail.com wrote:
Hello!
How can I do 100 simulations of length 17 from bivariate bivariate normal
distribution, if I know all 5 parameters?
--
View
The gls function in the nlme package is one approach.
If you know the covariance matrix exactly (it is just numerical with
nothing that needs to be estimated) then you can also take the Cholesky
decomposition of the inverse of the covariance matrix (or other square root
method) and multiply the x
How to change a factor to numeric is FAQ 7.10
Also look at the `colClasses` argument to read.table/read.csv for a way to
specify the type of the column when you are reading it in.
On Mon, Nov 26, 2012 at 8:59 AM, EcoFranc rob...@gmx.de wrote:
Hi,
maybe somebody would be kind enough to help
Duncan showed the way to do it in 1 plotting device, but it may be easier
to just open 2 graphics devices, arrange them side by side, then set up the
code to let you select in one, then switch to the other to do the plotting.
See ?dev.new and ?dev.set.
On Tue, Nov 27, 2012 at 8:46 AM, Baoqiang
Yes, I meant FAQ 7.21, must have stuttered in typing, but it is always good
to read other FAQs while looking for a specific one.
I did read your full description, though whether I fully understand or not
is yet to be seen.
It seems like a lot of what you want to do could be simplified by using
In addition, if you go the route of a data frame then the functions to look
at are tapply, aggregate, and ave.
On Wed, Nov 28, 2012 at 2:02 PM, Greg Snow 538...@gmail.com wrote:
Yes, I meant FAQ 7.21, must have stuttered in typing, but it is always
good to read other FAQs while looking
You have all the information that you need to do the pooled t-test using
the formula in many intro stats textbooks and probably on wikipedia as
well, just plug your numbers into the formulas in the book (R can act as
the calculator to make this easier).
Or sometimes simpler (for the human, the
Look at the updateusr function in the TeachingDemos package. One of the
examples shows adding lines to a barplot.
On Fri, Nov 23, 2012 at 12:33 AM, Ripples 9ripp...@gmail.com wrote:
Hi,
I'm trying to plot stacked barplot with lines on it. Here is the data.
emp days val1 val2 score
1 21
Why are you using jpeg to create files with a .png extension? wouldn't it
make more sense to use the png function?
There are a couple of options, you could plot using the matplot function
(with different subsets in the loop).
Or you can start an initial empty plot and use lines to add to it,
The key is to not change the margins, set them once and stick with those
margins. The next question then becomes how do I leave area at the
top/bottom for the title and common axis? to which the answer is Set
outer margins at the beginning.
Modifying your code:
y-rnorm(1:100)
x-rnorm(1:100)
Your code works for me, can you tell us what output you are getting, what
output you expect to see, and how they differ?
On Sat, Nov 10, 2012 at 11:23 AM, parvez_200207 parvez.r...@uef.fi wrote:
hi
could you help me to solve this issue
Question:
Using command rweibull(100,8,15), simulate n
This is to all R-helpers (Sarah is just the one that I am replying to),
Have we become a little too draconian on the not a homework help list
issue?
Now if someone just states the HW question, gives no indication that they
have done anything to try to solve it themselves, and expects us to give
There are still a lot of things that you did not specify, but maybe this
will get you started:
n.groups - sample(3:6, 1)
n.species - sample(2:10, n.groups, replace=TRUE)
n.animals - sample( 10:25, sum(n.species), replace=TRUE )
mu.g - rnorm(n.groups, 100, 50)
mu.species - rnorm( sum(n.species),
Others have shown you ways to do what you asked, and what you asked happens
to also be FAQ 7.22 (but with terms different enough that that FAQ is not
obvious).
However the solutions that you are tending towards are running into the
problems addressed in fortune(106) and fortune(236), the
The Dirichlet distribution will give values that sum to one and with the
correct conditions will have uniform margins. There are multiple packages
that will generate data from Dirichlet distributions, either gtools or
gmodels is one of them, but there are others as well.
On Wed, Nov 7, 2012 at
Yes, it is most likely due to scoping. It is safest to create a data
frame with all the data in it, then pass that to the data argument of
lm.
On Thu, Nov 1, 2012 at 2:25 AM, Eva Prieto Castro evapcas...@yahoo.es wrote:
Hi,
I have a problem in relation with a packahe I made. It runs on my
, Greg Snow 538...@gmail.com wrote:
Yes, it is most likely due to scoping. It is safest to create a data
frame with all the data in it, then pass that to the data argument of
lm.
On Thu, Nov 1, 2012 at 2:25 AM, Eva Prieto Castro evapcas...@yahoo.es
wrote:
Hi,
I have a problem in relation
It is not clear exactly what you want.
Do you want the final object to be a matrix with 2 columns (or the
number of columns equal to the number of items in the list? If so you
could just use do.call(cbind, x) and you would not need reduce or
paste (paste is the wrong tool unless you want a
First your response in the formula is a matrix which causes the lm
function to return an object of type 'mlm' for multivariate linear
model. Then when you run the stepAIC function it runs the addterm
function which looks for a method(function) to add terms to mlm
objects. However nobody has
You can use the lmList function in the nlme package to do several
seperate regressions, or use a model that allows for multiple
intercepts. Possibly Xvalues ~ 0 + log(Qvalues)*Tfac or Xvalues ~ 0 +
Tfac + log(Qvalues):Tfac (assuming Tfac is a factor).
On Tue, Oct 30, 2012 at 5:53 AM, Alex van
On Tue, Oct 30, 2012 at 8:00 AM, Bart Joosen bartjoo...@hotmail.com wrote:
Take a look at ?assign
Then read fortune(236)
Then learn about lists and environments to learn better methods.
Bart, without context it is harder for use to contribute useful advice.
Bart
--
View this message
301 - 400 of 2156 matches
Mail list logo