Christoph Scherber agr.uni-goettingen.de> writes:
>
> Dear Henric,
>
> Thanks, now it works; but how reliable are these estimates? Especially
> with p-values close to 0.05 it is of course important that the range of
> the estimates is not too large. I´ve just run several simulations, each
> o
Wensui Liu gmail.com> writes:
>
> chris,
> if you could create a package similar to proc sql in SAs, that will be
> so sweet.
Nice task, but I would not suggest to go this way. I found it quite easy to
work with R subsetting system after you get some practice and that is more
than enough - at
Hello R-lists,
I have created an extension library called "Rdonlp2".
This is a wrapper for Peter Spellucci's DONLP2:
http://plato.la.asu.edu/donlp2.html
DONLP2 is a standalone C program that can solve
following minimization problem:
min f(x) subject to
* parameter bounds
* linear (equality) cons
Why do you need to use a data frame? A list will give you the flexibility
you want:
d <- list( x=list( c(1,2), c(5,2), c(9,1) ), y=c( 1, -1, -1) )
Then you can access the individual elements
> d$x
[[1]]
[1] 1 2
[[2]]
[1] 5 2
[[3]]
[1] 9 1
> d$y
[1] 1 -1 -1
> d$x[[1]]
[1] 1 2
-Christos
Hi there,
I had a serious problem here . Consider the following Bayesian
model(discretized variance gamma):
#Likelihood
J[i]<-lambda*G[i]+sigma*sqrt(G[i])*rnorm(0,1)
G[i]<-rgamma(1/nu,1/nu)
#Prior:
nu<-rinvgamma(m,M)
# Parameters
lambda=-.04 ; sigam=.38; nu=6.4
I'd like to have a data.frame structured something like the following:
d <- data.frame (
x=list( c(1,2), c(5,2), c(9,1) ),
y=c( 1, -1, -1)
)
The reason is this: 'd' is the training data for a machine learning
algorithm. d$x is the independent data, and d$y is the dependent
data.
In genera
See ?tryCatch. /Henrik
On 2/12/07, Stephen Bond <[EMAIL PROTECTED]> wrote:
> Could smb please help with try-catch encapsulating a function for
> downloading. Let's say I have a character vector of symbols and want to
> download each one and surround by try and catch to be safe
>
> # get.hist.quote
On Feb 12, 2007, at 6:28 PM, Paul Lynch wrote:
> I'm trying to build R on RedHat EL4. The compile went fine, but a
> make check ran into a problem and produced a file
> "internet.Rout.fail". Judging by the last part of that file, it was
> trying to run an R routine called "httpget" to retrieve t
I'm trying to build R on RedHat EL4. The compile went fine, but a
make check ran into a problem and produced a file
"internet.Rout.fail". Judging by the last part of that file, it was
trying to run an R routine called "httpget" to retrieve the URL
http://www.stats.ox.ac.uk/pub/datasets/csb/ch11b.
Could smb please help with try-catch encapsulating a function for
downloading. Let's say I have a character vector of symbols and want to
download each one and surround by try and catch to be safe
# get.hist.quote() is in library(tseries), but the question does not
depend on it, I could be sour
Given the response by Carles Berry, you should probably really think
about what you want to do with the results (I'm hoping that you do not
plan to look at every R^2 value personally). For instance if you want
to find which variable gives the highest R^2 value for each variable,
then this approach
One approach is to split up the work of doing the correlations, if you
give the 'cor' function 2 matricies then it gives you the correlations
between all pairs of columns. Since you said it works fine with 10,000
columns but not 30,000 you could split into 3 pieces and do something
like (untested)
On Mon, 12 Feb 2007, andy1983 wrote:
>
> I would like to compare every column in my matrix with every other column and
> get the r-squared.
>
> I tried using the following formula and looping through every column:
>> summary(lm(matrix[,x]~matrix[,y]))$r.squared
> If I have 10,000 columns, the loop
We don't know how much memory your computer has and how exactly you plan
to generate these samples, so no one can really help you all that much.
Its best for you to just code up what you want to do and find out if you
get any memory errors or not. If you run into trouble and you can show
us exactl
I would like to compare every column in my matrix with every other column and
get the r-squared.
I tried using the following formula and looping through every column:
> summary(lm(matrix[,x]~matrix[,y]))$r.squared
If I have 10,000 columns, the loops (10,000 * 10,000) take forever even if
there is
...maybe library(neural) with examples should a good staring point.
regards, christian
> I am interested in Neural network models in R. Is
> there any reference material/tutorial which i can use.
> Regards,
>
>
>
> ___
Jim Lemon wrote:
> Xiaohui wrote:
>> ... Then, I did a heatmap for 'test' matrix. But for now, I want to
>> specify each of the cell in the heatmap according to the values of
>> the corresponding matrix elements of test.
>>
>> Let's say: col<-c("red","yellow","green")
>>
>> for test[1,1], the col
Hi
Randall C Johnson [Contr.] wrote:
> Hello,
> I'm trying to determine the width of a plotting point (in inches) in the
> grid package. I naively thought I could create a pointsGrob with only one
> point and get the width (as tried below), but this results in an object with
> a size of 0inches (
the one I will recommend is MASS by Dr B. Ripley.
On 2/12/07, vinod gullu <[EMAIL PROTECTED]> wrote:
> I am interested in Neural network models in R. Is
> there any reference material/tutorial which i can use.
> Regards,
>
>
>
> _
I am interested in Neural network models in R. Is
there any reference material/tutorial which i can use.
Regards,
TV dinner still cooling?
Check out "Tonight's Picks" on Yahoo! TV.
__
On 2/12/2007 10:20 AM, Ben Fairbank wrote:
> When signing off R or trying to save a workspace in Windows XP pro SP2,
> I receive the following error message -
>
>
>
> save.image("C:\\Program Files\\R\\R-2.4.1\\Responses3.RData")
>
>
>
> Error in save.image("C:\\Program Files\\R\\R-2.4.1\\Re
For your outlier question, look at the 'range' argument to the boxplot
function. It defaults to 1.5 meaning that any points more than 1.5*IQR
from the 1st and 3rd quartiles are considered outliers. If you make
this smaller, you will potentially see more outliers, if you make it
larger then it wil
An old rule of thumb was that you should have 6 times as much memory as
your dataset will take. But I think pretty much everything has been
improved since then, so you should be able to get by with less (others
may be able to give a better rule of thumb these days).
You might want to look at the
Hi Everyone,
I often use the 'safe prediction' feature available through glm().
Now, I'm at a situation where I must use biglm:::bigglm.
## begin example
library(splines)
library(biglm)
ff <- log(Volume)~ns(log(Girth), df=5)
fit.glm <- glm(ff, data=trees)
fit.biglm <- bigglm(ff, data=trees)
pr
Dear All,
I am planning to run a Monte-Carlo experiment which involves to do roughly
100.000 times the following
1- Generating a sample of, say, 50.000 numbers from an ARMA or GARCH
2- Doing some regressions on the series
3- On each regression storing one special value from the results into a
s
Greetings,
I previously asked if I could link R to Microsoft Server / Client.
The answer has been pouring in from around the world.
The answer is "Yes" I'm now exploring RODBC and learning the ropes.
Thanks to everyone who replied.
--
Steve Friedman
Computational Ecology and Visulation Labor
Hello,
I'm trying to determine the width of a plotting point (in inches) in the
grid package. I naively thought I could create a pointsGrob with only one
point and get the width (as tried below), but this results in an object with
a size of 0inches (changing cex has no effect). Does anyone have a b
Yes, you can use RODBC to connect to MS SQL. You can query, issue
commands, and save to the db from R and it works very well (at least for
me).
For example, with a database named xf, create an ODBC data source and
then link open a connection in R:
library(RODBC)
xf <- odbcConnect("xf", username,
Okay
First- I apologise for my poor use of terminology (error vs. warning).
Second- thanks for pointing out what I hadn't noticed before- when I pass
one case for selection to subset, I get all observations. When I pass two
cases (as a vector), I get every second case for both cases (if both a
When signing off R or trying to save a workspace in Windows XP pro SP2,
I receive the following error message -
save.image("C:\\Program Files\\R\\R-2.4.1\\Responses3.RData")
Error in save.image("C:\\Program Files\\R\\R-2.4.1\\Responses3.RData") :
recursive default argument referen
chris,
if you could create a package similar to proc sql in SAs, that will be so sweet.
On 2/12/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
> Hi R-users,
>
> This note will interest people who would like to use sql statements on R data
> frames (a bit like proc sql in SAS). Please reply to
The solution to most "factors" questions on the R mailing list is
to set the global option stringsAsFactors to F. Make it your part of your
default R startup. Even better, do what we have done at Mayo for the
last 10+ years and make it the default for your whole unit. (150+ users,
20+ years of
Dear Steve,
Reading the data shouldn't be a problem with RODBC. I've been doing that
plenty of times. I'm not sure about writing, but I suppose you probably
will.
Cheers,
Thierry
ir. Thierry Onkelinx
Instituut voor
steve,
i think you can use R to link to sql server directly with RODBC. but
it is not wise to dump the whole table from db into R and then do
manipulation, which will slow the speed of R.
On 2/12/07, Steve Friedman <[EMAIL PROTECTED]> wrote:
> Hello
>
> My colleagues and I have recently establish
Hi,
We are looking for a skilled statistician/programmer
seeking a career at the intersection of
finance, applied statistics and computer science.
This position requires a person with a strong background in
- data analysis,
- design and implementation of algorithms,
- software development
-
Hello
My colleagues and I have recently established a large database (40 tables
each with greater than 15 variables) in Microsoft's SQL Server 2000.
Currently we are accessing this database via SQL client running an Windows
XP. Our objectives are many fold including running SQL applications,
out
Dear Henric,
Thanks, now it works; but how reliable are these estimates? Especially
with p-values close to 0.05 it is of course important that the range of
the estimates is not too large. I´ve just run several simulations, each
of which yielding sometimes quite different p-values.
Best wishes
Does any package available in R contain the V-fold cross-validation method for
determining the right number of clusters?
Any help greatly appreciated
Kind regards
Dr Graham Leask
Economics and Strategy Group
Aston Business School
Aston University
Aston Triangle
Birmingham
B4 7ET
Tel: Di
Is the Duda & Hart clustering index available within R?
If so I would greatly appreciate the script file.
Many thanks
Kind regards
Dr Graham Leask
Economics and Strategy Group
Aston Business School
Aston University
Aston Triangle
Birmingham
B4 7ET
Tel: Direct line 0121 204 3150
ema
Den Må, 2007-02-12, 13:58 skrev Christoph Scherber:
> Dear all,
>
> I am currently analyzing count data from a hierarchical design, and I?ve
> tried to follow the suggestions for a correct estimation of p-values as
> discusssed at R-Wiki
> (http://wiki.r-project.org/rwiki/doku.php?id=guides:lmer-te
Dear all,
I am currently analyzing count data from a hierarchical design, and I?ve
tried to follow the suggestions for a correct estimation of p-values as
discusssed at R-Wiki
(http://wiki.r-project.org/rwiki/doku.php?id=guides:lmer-tests&s=lme%20and%20aov).
However, I have the problem that my
Hello,
Below a modified function for displaying a correlation matrix by Vincent
Zoonekynd posted in his excellent tutorial. Compared to the matrix and the
legend, the colouring of the polygons is wrong. The error is in the
sequence: col=col[N*(mat[i,j]+1)/2].
I cannot seem to devise a sequence
On 2/12/2007 2:50 AM, Patrick Connolly wrote:
> On Wed, 07-Feb-2007 at 07:07PM +1100, Jim Lemon wrote:
>
> |> Matthew Keller wrote:
> |> > Far from flaming you, I think you made a good point - one that I
> |> > imagine most people who use R have come across. The name "R" is a big
> |> > impedim
--- Gregor Gorjanc <[EMAIL PROTECTED]>
wrote:
> Jim Lemon bitwrit.com.au> writes:
> > Hi John,
> > You might have a look at "toNA" in the prettyR
> package. Wait for version
> > 1.0-4, just uploaded, as I have fixed a bug in
> that function.
>
> There is also a set of generic functions exactly
Ndoye Souleymane wrote:
> Hi,
>
> Let me suggest you to save your spss file in txt and use the read.table
> function to load your file in R.
> That is what I use to do.
The problem here is that the files are old data that were made with an ancient
version of spss and cannot be changed to txt (o
Hi,
I'm trying to speed up some code by using the Matrix package. For
most matrix manipulations I get a good increase in speed, however
applying the kronecker function is orders of magnitude slower than
with simple martix classes. In addition if I loop through kronecker
products for prof
plot(raw=TRUE) does this for you: there is little point in fitting by
orthogonal polynomials and then converting back.
As the help page says quite explicitly, poly() does not store the
orthogonal polynomials: please also consult the reference it gives. It
also points out that predict() creates
Hi Folks!
Im using the function poly to generate orthogonal polynomials, but Id like
to see the actual polynomials so that I could convert it to a polynomial
in my original variable. Is that possible and if so how do I do it?
/E
__
R-help@stat.math.eth
Hi R-users,
This note will interest people who would like to use sql statements on R data
frames (a bit like proc sql in SAS). Please reply to my only, unless you really
want to keep the entire R-help list posted on this.
I've been thinking about a packgage implementing sql queries in R. I'm a
Jim Lemon bitwrit.com.au> writes:
> Hi John,
> You might have a look at "toNA" in the prettyR package. Wait for version
> 1.0-4, just uploaded, as I have fixed a bug in that function.
There is also a set of generic functions exactly for such cases: unknownToNA(),
NAToUnknown() and isUnknown() in
Hi
as documentation does not say anything about changing quantile
computation it seems to me that you have 2 options:
1. change source code
2. change values computed by boxplot.stats and do plotting with
newly computed values by bxp
see respective help pages and look at structure for
51 matches
Mail list logo