This could be a bug in the e1071 svm code, but maybe not -- I guess I'll
send it here anyway. It's reproducible.
x <- seq (0.1,5,by=0.05)
y <- log(x) + rnorm (x, sd=0.2)
library(e1071)
m <- svm (x,y)
Process R segmentation fault at Tue Nov 20 23:34:19 2001
version
_
platform i686-pc-linux-gnu
arch i686
os linux-gnu
system i686, linux-gnu
status
major 1
minor 3.1
year 2001
month 08
day 31
language R
I used the Mandrake 8.0 RPM binaries from cran.r-project.org first, then
recompiled R myself and reinstalled e1071. Same results.
thanks,
--J.
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !) To: r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
jan> This could be a bug in the e1071 svm code, but maybe
jan> not -- I guess I'll send it here anyway. It's
jan> reproducible.
not really reproducible:
1) If you use random data (which is perfect),
you need an explicit set.seed(.) statement to make it reproducible
2) I can evaluate the code below (intel-Linux as you)
more than 100 times without a problem.
Martin
>> x <- seq (0.1,5,by=0.05)
>> y <- log(x) + rnorm (x, sd=0.2)
>> library(e1071)
>> m <- svm (x,y)
jan> Process R segmentation fault at Tue Nov 20 23:34:19 2001
>> version
jan> _
jan> platform i686-pc-linux-gnu
jan> arch i686
jan> os linux-gnu
jan> system i686, linux-gnu
jan> status
jan> major 1
jan> minor 3.1
jan> year 2001
jan> month 08
jan> day 31
jan> language R
jan> I used the Mandrake 8.0 RPM binaries from
jan> cran.r-project.org first, then recompiled R myself and
jan> reinstalled e1071. Same results.
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !) To: r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
This could be a bug in the e1071 svm code, but maybe not -- I guess I'll
send it here anyway. It's reproducible.
Which version of e1071 are you using (Current: 1.2-1)? This package had
a bug fix release one month ago.
-d
x <- seq (0.1,5,by=0.05)
y <- log(x) + rnorm (x, sd=0.2)
library(e1071)
m <- svm (x,y)
Process R segmentation fault at Tue Nov 20 23:34:19 2001
version
_
platform i686-pc-linux-gnu
arch i686
os linux-gnu
system i686, linux-gnu
status
major 1
minor 3.1
year 2001
month 08
day 31
language R
I used the Mandrake 8.0 RPM binaries from cran.r-project.org first, then
recompiled R myself and reinstalled e1071. Same results.
thanks,
--J.
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !) To: r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
Mag. David Meyer Wiedner Hauptstrasse 8-10
Vienna University of Technology A-1040 Vienna/AUSTRIA
Department for Tel.: (+431) 58801/10772
Statistics and Probability Theory mail: david.meyer@ci.tuwien.ac.at
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !) To: r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
"David" == David Meyer <david.meyer@ci.tuwien.ac.at> writes:
David> jan@rychter.com wrote:
>>
>> This could be a bug in the e1071 svm code, but maybe not -- I guess
>> I'll send it here anyway. It's reproducible.
David> Which version of e1071 are you using (Current: 1.2-1)? This
David> package had a bug fix release one month ago.
e1071_1.2-1.tar.gz
--J.
> x <- seq (0.1,5,by=0.05) y <- log(x) + rnorm (x, sd=0.2)
> library(e1071) m <- svm (x,y)
>>
>> Process R segmentation fault at Tue Nov 20 23:34:19 2001
>>
> version
>> _ platform i686-pc-linux-gnu arch i686 os linux-gnu system i686,
>> linux-gnu status major 1 minor 3.1 year 2001 month 08 day 31
>> language R
>>
>> I used the Mandrake 8.0 RPM binaries from cran.r-project.org first,
>> then recompiled R myself and reinstalled e1071. Same results.
>>
[...]
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !) To: r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
"David" == David Meyer <david.meyer@ci.tuwien.ac.at> writes:
David> jan@rychter.com wrote:
>>
>> This could be a bug in the e1071 svm code, but maybe not -- I guess
>> I'll send it here anyway. It's reproducible.
David> Which version of e1071 are you using (Current: 1.2-1)? This
David> package had a bug fix release one month ago.
To follow up on this (yes, I'm using 1.2-1), here is the backtrace. Hope
it helps.
library(e1071)
x <- seq (0.1,5,by=0.05)
y <- log(x) + rnorm (x, sd=0.2)
m <- svm (x,y)
Program received signal SIGSEGV, Segmentation fault.
0x4011ffe8 in realloc () from /lib/libc.so.6
(gdb) bt
#0 0x4011ffe8 in realloc () from /lib/libc.so.6
#1 0x403859ea in Cache::get_data (this=0x8b01d00, index=145759521, data=0xbfffd9b4,
len=99) at svm.cpp:126
#2 0x4038b239 in SVR_Q::get_Q (this=0xbfffdb00, i=100, len=198) at svm.cpp:1223
#3 0x4038658d in Solver::Solve (this=0xbfffdb50, l=198, Q=@0xbfffdb00, b_=0x8b01148,
y_=0x8b01780 '\001' <repeats 99 times>, 'ÿ' <repeats 99 times>, alpha_=0x8b00b10,
Cp=1, Cn=1, eps=0.001, si=0xbfffdc20, shrinking=1) at svm.cpp:534
#4 0x40388249 in solve_epsilon_svr (prob=0xbfffde40, param=0xbfffde50, alpha=0x8b007f0,
si=0xbfffdc20) at svm.cpp:1195
#5 0x40388aac in svm_train_one (prob=0xbfffde40, param=0xbfffde50, Cp=0, Cn=0)
at svm.cpp:1505
#6 0x40388c54 in svm_train (prob=0xbfffde40, param=0xbfffde50) at svm.cpp:1584
#7 0x40384202 in svmtrain () from /local/lib/R/library/e1071/libs/e1071.so
#8 0x0808ccd9 in do_dotCode ()
#9 0x08093e23 in Rf_eval ()
#10 0x0809546a in do_set ()
#11 0x08093c6e in Rf_eval ()
#12 0x08094db9 in do_begin ()
#13 0x08093c6e in Rf_eval ()
#14 0x08094238 in Rf_applyClosure ()
#15 0x080b423b in applyMethod ()
#16 0x080b47ce in Rf_usemethod ()
#17 0x080b49dd in do_usemethod ()
#18 0x08093c6e in Rf_eval ()
#19 0x08094238 in Rf_applyClosure ()
#20 0x08093eaa in Rf_eval ()
#21 0x0809546a in do_set ()
#22 0x08093c6e in Rf_eval ()
#23 0x080a906b in R_ReplConsole ()
#24 0x080a980a in run_Rmainloop ()
#25 0x080ff8e8 in main ()
#26 0x400bb5b0 in __libc_start_main () from /lib/libc.so.6
(gdb)
--Jan
> x <- seq (0.1,5,by=0.05) y <- log(x) + rnorm (x, sd=0.2)
> library(e1071) m <- svm (x,y)
>>
>> Process R segmentation fault at Tue Nov 20 23:34:19 2001
>>
> version
>> _ platform i686-pc-linux-gnu arch i686 os linux-gnu system i686,
>> linux-gnu status major 1 minor 3.1 year 2001 month 08 day 31
>> language R
>>
>> I used the Mandrake 8.0 RPM binaries from cran.r-project.org first,
>> then recompiled R myself and reinstalled e1071. Same results.
>>
>> thanks, --J.
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !) To: r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._