similar to: Problem tuning an SVM

Hi there - I have been trying to generate some simple stats and save the results to a file. My data looks like this: x y z exp 0 3 5 1 2 11 10 1 4 4 5 1 7 6 4 1 11 1 2 2 5 7 1 2 3 3

Hi there - Is there a way to write statistical models (trees, naïve Bayes, SVM, etc) to a file and import them again without loss of information? Thanks - Ton Ton van Daelen, PhD Director, Application Support Tel: (858) 279-8800 ext 217 Fax: (858) 279-8804 Web: www.scitegic.com Register now for the 2004 Pipeline Pilot user group meeting Jan 28-30 in San Diego:

Hi all - I am trying to tune an SVM model by optimizing the cross-validation accuracy. Maximizing this value doesn't necessarily seem to minimize the number of misclassifications. Can anyone tell me how the cross-validation accuracy is defined? In the output below, for example, cross-validation accuracy is 92.2%, while the number of correctly classified samples is (1476+170)/(1476+170+4) =

Dear All the problem I have is as follows. I have attribute data in a number of data.frames identified 1:58 (the column Null is just there to stop it becoming a list, which caused me trouble, the data I am interested in is in column 1, although other data frames have multiple categories/cols). The attribute data frames are always organised with 58 rows. ID Null 1 NA 1

Hello all, Sorry for this email not directly related to R developement. I just come from a nice demonstration session from Scitegic about their Pipeline Pilot (PP) software, and especially their 'R collection' which brings R calculations into the software (http://www.scitegic.com/documents/RStats_Collection.pdf). I looked carefully on the way they do it: they pass data from PP to R

I'm looking for an efficient way of removing zero-variance columns from a large matrix. Any suggestions? Thanks, - Moises [[alternative HTML version deleted]]

Hi: I am a newer in using R for data mining, and find the "e1071" pakage an excellent tool in doing data mining work! what frustrated me recently is that when I using the function "svm" and using the "cross=10" parameters, I got all the "accuracies" of the model greater than 1. Isn't that the accuracy should be smaller than 1? so I wander how, the

I'm using the following command to run R in Windows "Rterm --no-save --no-restore < "Rscriptfile" > "Rstdoutfile" How can I capture the text sent by R to stderr in a file? Thanks, - Moises [[alternative HTML version deleted]]

Dear All, It's been several times I observed a very strange behaviour with R 1.2.0 under Windows NT 4.0 (with sp5): accuracies are lost in many cases. Things are back to normal when I exit R and restart it. Here is an example of my last encounter with this: > bt1 <- c(7.3, 8.1, 8.4, 7, 6, 7.2, 6.7, 8.3, 9, 7.4, 8.6, 6.2, 6.9, 7.6, 5.3, 5.6) > bt1 [1] 7 8 8 7 6 7 6 8 9 7 8 6 6 7 5 5

I am implementing an image classification algorithm using the randomForest package. The training data consists of 31000+ training cases over 26 variables, plus one factor predictor variable (the training class). The main issue I am encountering is very low overall classification accuracy (a lot of confusion between classes). However, I know from other classifications (including a regular decision

Hi, I am using lmer (from the package lme4) to predict a binary response variable (REL) from a bunch of fixed effects and two random effects (Speaker_ID and NPhead_lemma): fit <- lmer(REL ~ SPEAKER_GENDER + log(SPEECHRATE) + SQSPEECHRATE + ..... + (1|Speaker_ID) + (1|NPhead_lemma), family="binomial", data=data.lmer, method="Laplace", model=T, x=T) I

Hi Duncan, I'm not sure about this: + if (!Accuracy) + // If it's not a floating point number then it must be 'fast'. + return getFastAccuracy(); Since you allow accuracies bigger than 1 in setFPAccuracy(), integers should be treated as float. Or at least assert. Also, I'm thinking you should carry the annotation forward on all uses of an annotated result, or make

I'm trying to install the VR package for version 1.9.1 but I'm getting the following error message: > install.packages("VR") trying URL `http://cran.r-project.org/bin/windows/contrib/1.9/PACKAGES' Content type `text/plain; charset=iso-8859-1' length 20716 bytes opened URL downloaded 20Kb trying URL

Hi, I'm a developer working on an application for Windows XP and have been using the Win32 API call FileLockEx() on shared files (writing or reading). This has been working fine, but we now have a customer where this causes the following error when trying to lock a file on a samba network drive (it works fine if the file is local or on a windows shared drive): LockFileEx returns false.

On Jan 27, 2012, at 3:40 PM, Devang Patel wrote: > [ removing cfe-dev from the cc list ] > > On Jan 27, 2012, at 1:31 PM, Dan Gohman wrote: > >> On Jan 27, 2012, at 11:20 AM, Devang Patel wrote: >> >>> >>> On Jan 26, 2012, at 2:10 PM, Dan Gohman wrote: >>> >>>> On Jan 26, 2012, at 12:54 PM, Devang Patel wrote:

Hi Renato, > I'm not sure about this: > > + if (!Accuracy) > + // If it's not a floating point number then it must be 'fast'. > + return getFastAccuracy(); > > Since you allow accuracies bigger than 1 in setFPAccuracy(), integers > should be treated as float. Or at least assert. the verifier checks that the accuracy operand is either a floating

Not sure if this document will be shown to any dtrace users but here is a link if you want to comment on it http://sources.redhat.com/systemtap/wiki/SystemtapDtraceComparison I have went line by line and commented on its content and in accuracies and posted it both to the systemtap email list and my blog, you can see the blogged version at

I'm using caret to assess classifier performance (and it's great!). However, I've found that my results differ between R2.* and R3.* - reported accuracies are reduced dramatically. I suspect that a code change to kernlab ksvm may be responsible (see version 5.16-24 here: http://cran.r-project.org/web/packages/caret/news.html). I get very different results between caret_5.15-61 +

As a follow up from my previous post, I am pondering some changes to my env. I read that with AR I should change the verification timeout to be lower than the MySQL db Interactive_timeout. Now, I have done some performancing tuning to mysql in LAMP env''s but not RoR. MySQL typical set the default interactive_timeout to 8hrs (28800) this has bitten me a few times with sleeping processess

I have been running a public MySQL server for over 4 years. The system was a 1GHz box with 256MB of RAM. MySQL puts each database into a seperate directory in a single "data" directory. Once the old system reached about 10K databases the connection times increased 10X or more. I attributed the speed problems to a possible filesystem limitation with a large number of files. That