svm plot question - R-help | R Mailing Lists

Fri, Dec 8, 2006 4:32 AM #

Aimin:

1) Please do not spam the r-help list---one request per issue (and two
private mails to the code author) really suffice. Not all contributors
to the R-project are on-line 24/24, and have time to provide immediate
answers.

2) The error occurs because plot.svm() currently does not set valid
defaults for categorical dimensions you are conditioning on for your
2D-plot (in your example: 'P' and 'Aa') which certainly is a bug. I will
commit a fix for the next release of e1071. For the time being, you will
have to explicitly specify the levels of 'P' and 'Aa':

plot(m.svm,p5.new,As~Cur, slice = list(P = factor("821p", levels =
levels(P)), Aa = factor("ALA", levels = levels(Aa))))

(Note that the defaults for the "slice" argument are completely
arbitrary anyway).

Thanks for pointing this out,

David

Aimin Yan wrote:

Dr. David Meyer
Department of Information Systems and Operations

Vienna University of Economics and Business Administration
Augasse 2-6, A-1090 Wien, Austria, Europe
Tel: +43-1-313 36 4393
Fax: +43-1-313 36 90 4393
HP:  http://wi.wu-wien.ac.at/~meyer/

Aimin Yan

Fri, Dec 8, 2006 11:27 AM #

thanks, I did get this plot.
Before I have this problem, I did get a plot by my code.
However after I change a little my code. it doesn't work.
It is pity not saving my original code.

Now the question is the plot I get using your code is different from
what I got before.
Moreover I did remember I use plot(m.svm,p5.new,As~Cur)

Do you know why?

Thanks,

Aimin

At 06:32 AM 12/8/2006, David Meyer wrote:

Aimin:

1) Please do not spam the r-help list---one request per issue (and two
private mails to the code author) really suffice. Not all contributors
to the R-project are on-line 24/24, and have time to provide immediate
answers.

2) The error occurs because plot.svm() currently does not set valid
defaults for categorical dimensions you are conditioning on for your
2D-plot (in your example: 'P' and 'Aa') which certainly is a bug. I will
commit a fix for the next release of e1071. For the time being, you will
have to explicitly specify the levels of 'P' and 'Aa':

plot(m.svm,p5.new,As~Cur, slice = list(P = factor("821p", levels =
levels(P)), Aa = factor("ALA", levels = levels(Aa))))

(Note that the defaults for the "slice" argument are completely
arbitrary anyway).

Thanks for pointing this out,

David

Aimin Yan wrote:

I have a question about svm in R

I run the following code, all other is ok,
but plot(m.svm,p5.new,As~Cur) is not ok

Do you know why?

install.packages("e1071")
library(e1071)
library(MASS)
p5 <- read.csv("http://www.public.iastate.edu/~aiminy/data/p_5_2.csv")
p5.new<-subset(p5,select=-Ms)
p5.new$Y<-factor(p5.new$Y)
levels(p5.new$Y) <- list(Out=c(1), In=c(0))
attach(p5.new)
m.svm<-svm(Y~P+Aa+As+Cur,data=p5.new)
summary(m.svm)
plot(m.svm,p5.new,As~Cur)

Here is output:

install.packages("e1071")

--- Please select a CRAN mirror for use in this session ---
trying URL

'http://rh-mirror.linux.iastate.edu/CRAN/bin/windows/contrib/2.4/e1071_1.5-16.zip'

Content type 'application/zip' length 592258 bytes
opened URL
downloaded 578Kb

package 'e1071' successfully unpacked and MD5 sums checked

The downloaded packages are in
        C:\Documents and Settings\aiminy\Local
Settings\Temp\RtmpY0B2qb\downloaded_packages
updating HTML package descriptions

library(e1071)

Loading required package: class

library(MASS)
p5 <- read.csv("http://www.public.iastate.edu/~aiminy/data/p_5_2.csv")
p5.new<-subset(p5,select=-Ms)
p5.new$Y<-factor(p5.new$Y)
levels(p5.new$Y) <- list(Out=c(1), In=c(0))
attach(p5.new)
m.svm<-svm(Y~P+Aa+As+Cur,data=p5.new)
summary(m.svm)

Call:
svm(formula = Y ~ P + Aa + As + Cur, data = p5.new)


Parameters:
   SVM-Type:  C-classification
 SVM-Kernel:  radial
       cost:  1
      gamma:  0.04

Number of Support Vectors:  758

 ( 382 376 )


Number of Classes:  2

Levels:
 Out In

plot(m.svm,p5.new,As~Cur)

Error in scale(newdata[, object$scaled, drop = FALSE], center =
object$x.scale$"scaled:center",  :
        (subscript) logical subscript too long

--
Dr. David Meyer
Department of Information Systems and Operations

Vienna University of Economics and Business Administration
Augasse 2-6, A-1090 Wien, Austria, Europe
Tel: +43-1-313 36 4393
Fax: +43-1-313 36 90 4393
HP:  http://wi.wu-wien.ac.at/~meyer/

David Meyer

Sat, Dec 9, 2006 2:06 AM #

Aimin:

hard to tell. IMO, without specifying defaults, it could only work with
purely numeric data since factors were wrongly processed.

David.

Aimin Yan wrote:

thanks, I did get this plot.
Before I have this problem, I did get a plot by my code.
However after I change a little my code. it doesn't work.
It is pity not saving my original code.

Now the question is the plot I get using your code is different from
what I got before.
Moreover I did remember I use plot(m.svm,p5.new,As~Cur)

Do you know why?

Thanks,

Aimin

At 06:32 AM 12/8/2006, David Meyer wrote:

Aimin:

1) Please do not spam the r-help list---one request per issue (and two
private mails to the code author) really suffice. Not all contributors
to the R-project are on-line 24/24, and have time to provide immediate
answers.

2) The error occurs because plot.svm() currently does not set valid
defaults for categorical dimensions you are conditioning on for your
2D-plot (in your example: 'P' and 'Aa') which certainly is a bug. I will
commit a fix for the next release of e1071. For the time being, you will
have to explicitly specify the levels of 'P' and 'Aa':

plot(m.svm,p5.new,As~Cur, slice = list(P = factor("821p", levels =
levels(P)), Aa = factor("ALA", levels = levels(Aa))))

(Note that the defaults for the "slice" argument are completely
arbitrary anyway).

Thanks for pointing this out,

David

Aimin Yan wrote:

I have a question about svm in R

I run the following code, all other is ok,
but plot(m.svm,p5.new,As~Cur) is not ok

Do you know why?

install.packages("e1071")
library(e1071)
library(MASS)
p5 <- read.csv("http://www.public.iastate.edu/~aiminy/data/p_5_2.csv")
p5.new<-subset(p5,select=-Ms)
p5.new$Y<-factor(p5.new$Y)
levels(p5.new$Y) <- list(Out=c(1), In=c(0))
attach(p5.new)
m.svm<-svm(Y~P+Aa+As+Cur,data=p5.new)
summary(m.svm)
plot(m.svm,p5.new,As~Cur)

Here is output:

install.packages("e1071")

--- Please select a CRAN mirror for use in this session ---
trying URL

'http://rh-mirror.linux.iastate.edu/CRAN/bin/windows/contrib/2.4/e1071_1.5-16.zip'

Content type 'application/zip' length 592258 bytes
opened URL
downloaded 578Kb

package 'e1071' successfully unpacked and MD5 sums checked

The downloaded packages are in
        C:\Documents and Settings\aiminy\Local
Settings\Temp\RtmpY0B2qb\downloaded_packages
updating HTML package descriptions

library(e1071)

Loading required package: class

library(MASS)
p5 <- read.csv("http://www.public.iastate.edu/~aiminy/data/p_5_2.csv")
p5.new<-subset(p5,select=-Ms)
p5.new$Y<-factor(p5.new$Y)
levels(p5.new$Y) <- list(Out=c(1), In=c(0))
attach(p5.new)
m.svm<-svm(Y~P+Aa+As+Cur,data=p5.new)
summary(m.svm)

Call:
svm(formula = Y ~ P + Aa + As + Cur, data = p5.new)


Parameters:
   SVM-Type:  C-classification
 SVM-Kernel:  radial
       cost:  1
      gamma:  0.04

Number of Support Vectors:  758

 ( 382 376 )


Number of Classes:  2

Levels:
 Out In

plot(m.svm,p5.new,As~Cur)

Error in scale(newdata[, object$scaled, drop = FALSE], center =
object$x.scale$"scaled:center",  :
        (subscript) logical subscript too long

-- 
Dr. David Meyer
Department of Information Systems and Operations

Vienna University of Economics and Business Administration
Augasse 2-6, A-1090 Wien, Austria, Europe
Tel: +43-1-313 36 4393
Fax: +43-1-313 36 90 4393
HP:  http://wi.wu-wien.ac.at/~meyer/

Dr. David Meyer
Department of Information Systems and Operations

Vienna University of Economics and Business Administration
Augasse 2-6, A-1090 Wien, Austria, Europe
Tel: +43-1-313 36 4393
Fax: +43-1-313 36 90 4393
HP:  http://wi.wu-wien.ac.at/~meyer/