gmane.comp.lang.r.general

Re: Resampling a grid to coarsen its resolution

Sharpie — 2010-02-10T05:16:16


Steve Murray-2 wrote:

Check out interp.surface.grid() in the fields package.  I would also
recommend replacing your -9999 values with NAs  so they don't introduce
distortion on the edge of your grid during the interpolation.

prompts and running means

RagingJim — 2010-02-10T05:12:08

G'day, I am new user to R, and have been thrown in the deep end with a
something my company want me to write.

my code is as follows:

kenttemp=read.csv("mnowak.11.1.csv")
rows=nrow(kenttemp)-5
kent=kenttemp[1:rows,]              #have to remove the last 5 lines of the
graph as they interfere with rest of data
max(kent[,16],na.rm=TRUE)->ymax
ymax=ymax+200#This is to get vertical scale to fit more accurately
Ann=kent$Ann
kent=as.matrix(kent)
barplot(Ann,main="Annual Monthly Rainfail Data", xlab="Year",ylab="Rainfall
(mm)",ylim=c(0,ymax),col="blue",space=0,names.arg=c(kent[,2]),
cex.names=0.8)


All the code works fine, and does as I want it to, however I need to create
2 things.

Firstly, I need to somehow make it so that when the script is run (I realise
that there are no functions or anything yet, but that is still beyond my 2
day old knowledge of this program, and that will come as soon as I figure it
out), it will ask the user for the name of the csv file it wants to open,
and then prompt for what th

Re: The 'variables' attribute of terms()

Peter Ehlers — 2010-02-10T05:07:43

See Chapter 2 in the R Language Definition manual.

  -Peter Ehlers

Re: Bar plot

Peter Ehlers — 2010-02-10T04:50:53

Well, that makes sense. You can't put a title on a
plot that doesn't exist.

  -Peter

Re: Model matrix using dummy regressors or deviation regressors

2010-02-10T04:43:31

Instead of change the option globally, it is cleaner to pass the
contrast choice as an argument to aov(). It seems that 'contrasts.arg'
of model.matrix() can be used for this purpose.

I tried something like contrasts.args=c('contr.sum', 'contr.poly').
But it is not working. Would you please show me how to specify this
argument for the example that I gave in the original email?

Re: ggplot2 stacked line plot

Liam Blanckenberg — 2010-02-10T04:43:04

Hi Hadley,

Thank you for taking the time to help me with this - I've constructed
the following example to illustrate my problem:

require(ggplot2)

data.set <- data.frame(
 Time = c(rep(1, 4),rep(2, 4), rep(3, 4), rep(4, 4)),
 Type = rep(c('a', 'b', 'c', 'd'), 4),
 Value = c(10, 12, 14, 16, 14, 14, 20, 18, 18, 16, 22, 20, 26, 20, 24, 26)
)

When I use the following code:

   ggplot(data.set, aes(x = Time, y = Value, colour = Type)) +
geom_area(aes(fill = Type), position = 'stack')

I get a very nice stacked area chart - each layer on the chart
represents a Type, with the total height of all the layers
representing the total Value across all Types for a given Time period
- this is what you would get numerically with the code:

   with(data.set, aggregate(Value, by = list(Time), sum))

I want to achieve exactly the same thing as this, but instead of
having a coloured area to represent each layer I just want a coloured
line for each layer. When I use the following code to try and achieve
this:

   ggplot(data.

Re: OrdFacReg

Andrew Kosydar — 2010-02-10T04:33:36

Hi Dennis,

Thank you for your response.  No, NB is not a matrix, and I have no covariates.  Here's a very small sample of the data:

effectNB
  -0.0032001
  -0.1208003
  -0.0032002
  -7.6900001
  -1.4421002
  -0.0009001
  -0.0142003
  -5.0150000
  -0.0014002
  -0.0080003
  -2.3370002
  -0.0040501
  -0.1014001
  -0.0021000
  -0.0036002
  -0.0024003
  -1.1230001
  -0.0006002

I am purely interested in whether an increase in "NB" (for ex: from 0 to 1, or 3 to 4) predicts a directional change with "effect".

Any advise is greatly appreciated.

Many thanks,

Andrew

Re: transparent concentric circles

Jim Lemon — 2010-02-10T04:24:40

Hi Karin,
As Barry noted, you can use draw.circle in the plotrix package:

plot(0,xlim=c(-100,100),ylim=c(-100,100),type="n",axes=FALSE,
  main="Set intersection plot",xlab="",ylab="")
par(xpd=TRUE)
draw.circle(0,0,100,col="#ffccff")
draw.circle(0,0,75,col="#ff99ff")
draw.circle(0,0,50,col="#ff44ff")
draw.circle(0,0,25,col="#ff00ff")
par(xpd=FALSE)

These aren't transparent, but if you really need transparency, just add 
the alpha level to the colors. Look at boxed.labels to label the circles.

Jim

The 'variables' attribute of terms()

2010-02-10T04:20:09

I got the following output when I try the code at the end.

 language list(skips, Panel, Opening)
[1] "call"

I checked ?call, where 'call' is a function. Could somebody let me
know what a 'call' class is and what 'language' is?

##########
form1=skips ~ Panel * Opening
terms1=terms(form1)
str(attr(terms1, 'variables'))
class(attr(terms1, 'variables'))

Re: Bar plot

Sunita Patil — 2010-02-10T04:18:52

Hello Peter

Sorry I forgot to paste the error for tittle it was as follows

*Error in title(main = "Year Vs StudentsPassed") :   plot.new has not been
called yet*

Regards

Our Thoughts have the Power to Change our Destiny.
Sunita
Sent from Pune, MH, India

On Tue, Feb 9, 2010 at 11:22 PM, Peter Ehlers <ehlers< at >ucalgary.ca> wrote:


[[alternative HTML version deleted]]

estimators based on Truncated likelihood

2010-02-10T03:07:27


Dear Sir/Madam,

May I know if there is any function that estimates the ARCH or GARCH
models based on truncated likelihood?

Thanks,
Helen

Re: Fast way to determine number of lines in a file

kMan — 2010-02-10T03:59:52

It depends on the type of file and your system. 'count.fields()' is
impractical for large files because it generates a matrix with the same
number of dimensions as the file. It would be easier to use scan() with the
delimiter argument set up to read to the end of line marker, "\n" I believe,
and the 'what' argument set to a null list, so nothing is actually read.
Scan will still report the number of lines read. 

For flat files, and in windows, additional utilities installed with RTOOLS
(just need the tools-Cygwin dlls install) are the fastest that I know of. 

if(.Platform$OS.type=="windows"){ 
  system.time({ 
    cmd<-system(paste("/RTools/bin/wc -l","much_data.bin"), intern=TRUE) 
    cmd<-strsplit(cmd, " ")[[1]][1] 
    }) 
 }

Sincerely,
KeithC.

-----Original Message-----
From: Hadley Wickham [mailto:hadley< at >rice.edu] 
Sent: Monday, February 08, 2010 7:16 AM
To: R-help
Subject: [R] Fast way to determine number of lines in a file

Hi all,

Is there a fast way to determine the number of lines in a file?

Re: Resampling a grid to coarsen its resolution

Moshe Olshansky — 2010-02-10T03:55:47

One possibility I can see is to replace -9999 by NA and use mean with na.rm=TRUE.

--- On Wed, 10/2/10, Steve Murray <smurray444< at >hotmail.com> wrote:

Re: the hat ^ in regular expression

kMan — 2010-02-10T03:37:21

Use "\\^" instead.

For example:
df<-data.frame(c(1,2,4),c(4,3,2))
names(df)<-c("Amt","Resp")
df.form<-formula(Resp~0+Amt+I(Amt^2), data=df)
strsplit(as.character(df.form)[3], "\\^", perl=T)

should work just fine.

Sincerely,
KeithC.

-----Original Message-----
From: Gabor Grothendieck [mailto:ggrothendieck< at >gmail.com] 
Sent: Monday, February 08, 2010 6:01 AM
To: christophe dutang
Cc: r-help< at >r-project.org
Subject: Re: [R] the hat ^ in regular expression

Try this:

[1] "variable1" "variable2"

See http://gsubfn.googlecode.com for more info on strapply.

Another approach is

[1] "variable1" "variable2"


On Mon, Feb 8, 2010 at 6:22 AM, christophe dutang <dutangc< at >gmail.com> wrote:

looping problem

Roslina Zakaria — 2010-02-10T03:23:08

Hi R-users,
 
I have this code here:

library(numDeriv)
 
fprime <- function(z)
{ alp  <- 2.0165;
  rho  <- 0.868;
 
# simplified expressions
  a      <- alp-0.5
  c1     <- sqrt(pi)/(gamma(alp)*(1-rho)^alp)
  c2     <- sqrt(rho)/(1-rho)
  t1     <- exp(-z/(1-rho))
  t2     <- (z/(2*c2))^a
  bes1   <- besselI(z*c2,a)
  t1bes1 <- t1*bes1
  c1*t1bes1*t2
}
 
## Newton iteration
newton_gam <- function(z)
{ n   <- length(z)
  r   <- runif(n)
  tol <- 1E-6
  cdf <- vector(length=n, mode="numeric")
 
  for (i in 1:1000)
  { # numerical intergration to find the cdf
    cdf  <- integrate(fprime, lower = 0, upper = z)$value
 
    # Newton method
    z    <- z - (cdf - r)/fprime(z)
 
    if (tol < 1e-10) break
   }
  cbind(z,cdf)
}
 
bt1 <- 29.107 ; bt2 <- 41.517
x1  <- 30; x2 <- 10;
 
z <- (x1/bt1)+(x2/bt2); z
newton_gam(z)
 
[1] 1.271545
            z       cdf
[1,] 4.128138 0.6065616
 
But when I try with different values of X

write.zip?

spencerg — 2010-02-10T03:15:32

      Can one write a zip file from R? 


      I want to create a file with a name like "dat.zip", being a zip 
file containing "dat.csv".  I can create "dat.csv", then call 
"system('zip -r9 dat.zip dat.csv')".  Is there a better way? 


      I can use "gzfile" to write a gz file, but I don't know how to 
give that a structure that would support unzipping to a *.csv file. 


      Thanks,
      Spencer

estimators based on Truncated likelihood

2010-02-10T02:21:23

Dear Sir/Madam,

May I know if there is any function that estimates the ARCH or GARCH  
models based on truncated likelihood?

Thanks,
Helen

Re: using setMethod or setGeneric to change S4 accessor symbolfrom < at > to $

Markus Weisner — 2010-02-10T00:12:52

Thanks so much for your help.  I am realizing that I may be
over-complicating things for myself.  I have learned a ton about creating
methods, but I feel like I am trying to reinvent the data.frame class.
Basically, I am trying to create a data.frame type object where I can
enforce the header names and column data types.  I am trying to force the
user to setup the following fields:

   - event_number (character)
   - agency (factor)
   - unit_num (factor)
   - alarm (POSIXct)
   - priority (factor)

A user might use the following code:

event_number = c(1:5)
agency = c("CFD", rep("ACFR", 3), "CFD")
unit_num = c("E1", "T10", "E3", "E2", "BC1")
temp =  c("00:52:35", "06:58:18", "13:42:18", "20:59:45", "21:19:00")
alarm = as.POSIXct(strptime(temp, format="%H:%M:%S"))
priority = c("A", "E", "A", "C", "C")
data = data.frame(event_number=event_number, agency=agency,
unit_number=unit_num, alarm=alarm, priority=priority)

I have all sorts of functions that I am trying to incorporate into a package
for analyzing fire

Re: Double Integral Minimization Problem

Ravi Varadhan — 2010-02-10T02:57:05

Hi,

Is your code reproducible?

I can't even find the package "adapt" on the CRAN repository.  I am not sure what exactlt happened to that package, but do remember seeing something about it relatively recently in R-help.  Are you using an older version of adapt?

Ravi.

____________________________________________________________________

Ravi Varadhan, Ph.D.
Assistant Professor,
Division of Geriatric Medicine and Gerontology
School of Medicine
Johns Hopkins University

Ph. (410) 502-2619
email: rvaradhan< at >jhmi.edu


----- Original Message -----
From: MVika <mv56< at >st-andrews.ac.uk>
Date: Tuesday, February 9, 2010 5:41 pm
Subject: [R] Double Integral Minimization Problem
To: r-help< at >r-project.org

Re: simple subtraction in a single vector

jim holtman — 2010-02-10T02:46:42

Not exactly sure what you are asking: is this it

+ Tanks=c("a3","a4","c4","h4"))
  length Tanks
1      1    a3
2      2    a4
3      3    c4
4      4    h4
[1] 3


On Tue, Feb 9, 2010 at 9:37 PM, Marlin Keith Cox <marlinkcox< at >gmail.com> wrote:

Re: simple subtraction in a single vector

Peter Alspach — 2010-02-10T02:45:34

Tena koe Keith

?which

perhaps.  As in:

which(Tanks=='h4')-which(Tanks=='a3')

HTH ...

Peter Alspach