similar to: FBI Homicide data?

Dear R-listers, I would like to carry out a very basic descriptive analysis of homicides rates in Italy, taking into account both the spatial dimension (103 provinces) and the temporal dimension (10 years), but no covariates. In practice, what I would like to do is to describe spatio-temporal variation of homicide rates, identifying those combinations of province-year where the homicide rate

I'm running my script using mpirun -mp 4 and using snowfall+Rmpi on Linux 64bits. I receive the following error, but apparently without consequences on the results. Any idea? I'm able to reproduce it with a minimal script (below). Seems the critical issue is the for loop. Without it no error. Thanks for your help! TERM: Undefined variable. TERM: Undefined variable. TERM: Undefined

Sorry for wasting your time with an incorrect diagnosis in my previous mail! I'm running my script using mpirun -mp 4 and using snowfall+Rmpi on Linux 64bits. I receive the following error, that terminates the run. Any idea? I'm able to reproduce the problem with a minimal script (below). If n < 16371 it completes 9 iterations without problems. Already at 16371 it dies at iteration 7

I'm preparing to install 2.13, so I need a clarification about package location on disk. On window the packages I have installed are under C:\Users\mvalle\R\win-library\2.12 and the ones that come with R are under "C:\Program Files\R\library" If I execute update.packages() from the R gui, it fails because it finds e.g. Matrix as a package that should be updated but the process

I am hoping for some advice regarding limiting decimal points to 3. 'Round' produces the desired results except for the 97.5% confidence interval. Any advice as to how I modify the code to obtain output to 3 decimal points for all ouput is appreciated, regards Bob Green mod.multgran <-multinom(offence ~ grandiose * violent.convictions, data = kc, na.action = na.omit)

On Linux 2.13 by default links cairo, so png() uses it for rendering. One of my users has performance problem, so ask if it possible to temporary disable cairo during png() rendering so it reverts to grGraphics. My workaround for now is to configure R without cairo, but this affects all users that are forced to load and use CairoDevice if they wants better quality. I want to avoid this, is it

Any advice when to use denstity() and when the KernSmooth package bkde() to smooth a histogram? No specific problem to use either one, but I'm curious why there are two so similar implementations. Thanks! mario -- Ing. Mario Valle Data Analysis and Visualization Group | http://www.cscs.ch/~mvalle Swiss National Supercomputing Centre (CSCS) | Tel: +41 (91) 610.82.60 v.

Hi all! I have an ordered vector of values. The distribution of these values can be modeled by a sum of Gaussians. So I'm using the package 'mclust' to get the Gaussians's parameters for this 1D distribution. It works very well, but, for input sizes above 100.000 values it starts taking really forever. Unfortunately my dataset has around 4.6M values... My question: is it

Some time ago someone asked for things that make R difficult to master. Here I want to record one R behavior that took me off-guard yesterday. Moral of the story: vectorialize, but don't exaggerate. Hope it helps mario ### A very simple data frame tc <- textConnection( "prefix idx val A 1 11 A 2 22 B 1 33 B 2 44") s <- read.table(tc,

Good morning! Which is the preferred method to leave a sourced script and returning back to the '>' prompt? For example I search for certain files to be processed, but nothing should be done if they are not present. Normally I do in my script: f<-dir(pattern="qq") if(length(f) > 0) { ...process file list... } # script end But I don't like those {} encompassing

Whas missing the R in the command line: "mpirun -n --hostfile /home/hostfile R --no-save -f rtest.R" Hope this helps mario On 13-Jan-11 22:08, Justin Moriarty wrote: > Hi, > Just wanted to share a working example of doSNOW and foreach for an openMPI cluster. The function eddcmp() is just an example and returns some inocuous warnings. The

Hello, I don't know if the question pertains to Rmpi, snow or snowfall. I run my job by: mpirun -np N -hostfile $PBS_NODEFILE RMPISNOW -f my-script.r --slave In the snowfall sfInit call I have to specify one less CPU respect to the mpirun call sfInit(parallel=TRUE, cpus=N-1, type="MPI") otherwise I receive an error similar to: "cluster size N-1 already

n <- 63 a <- 1:n x <- a-1 y <- outer(x,a) matplot(x,y,type='l') Warning message: In matplot(x, y, type = "l") : default 'pch' is smaller than number of columns and hence recycled Why is it complaining if I specifically ask for type="l", so no pch involved? Annoyance or feature? The fix (if needed) is simple. In the matplot code change: if

Dear all, I'm using R 2.12.0 on Windows 7 (32bits) I created a filled contour from the attached data using the following code: load('bug.RData') pdf('bug.pdf', width=14, height=7) filled.contour(o4$x, o4$y, o4$z, color=rainbow, xlim=c(4,18), ylim=c(0,35), zlim=c(14,36)) dev.off() If you look (I used acrobat reader 9.4) at the attached bug.pdf, you see an incomplete grid of

[sorry, forgot some details...] I'm using optim(param, fun, method='L-BFGS-B', lower=lo, upper=up) to minimize a certain function. Often the minimization ends with the message: ERROR: ABNORMAL_TERMINATION_IN_LNSRCH What is optim() trying to say? What have I to change in my function to make the minimization succeed? Do you think using BBoptim() instead of optim() changes anything?

Hello! Suppose I have three charts like below. The top chart is a general overview and the bottom charts are related so some point of this chart. To make clear this relationship I want to draw a line between (4,0.9) in the top chart and (10,1) in the bottom-left one. Currently I add it manually using Inkscape on the resulting pdf file. Is it possible to add it inside R? Should I switch to other

I just upgraded my Win7 32bits installation to 2.14.0 after deinstalling 2.12.x First thing I moved the win-library from 2.12 to 2.14 and executed a update.packages(ask='graphics',checkBuilt=TRUE) (Swiss mirror). This aborts with the console message: Error in if (any(diff)) { : missing value where TRUE/FALSE needed And the antivirus (AVG) pops up complaining that colorspace.dll

Dear all, I have a multiline plot with each line labeled with a different letter. But I'm not able to make the legend display the same kind of pattern '-a-', instead the letter is overwritten by the line. A simpler legend with only the letter is not very visible and the pt.bg does nothing with letters. Any idea? plot(1:10,10:1,lty=1,type='b', lwd=2,pch='a')

Shouldn't -n be 4 in the bsub command? One master+3 slaves. This was required for snowfall, but I think doSNOW is similar. Hope it helps mario On 16-Dec-10 23:09, Marius Hofert wrote: > Dear expeRts, > > I try to get a minimal example for parallel computing via "foreach" + "doSNOW" to run on a computer cluster (Brutus from

Hi, I would love to make multiple histograms transposed one on another in order to show relation between the sets. I tried to write a function like this, but R tells me, that I cannot use add=FALSE in high-level commands. That's nice but I am supposed to do? rm(list=ls()) # what's wrong with underscore? #getwd("/home/matej/docs/skola/stat\_anal-cj3534/assign01/")