LIBSVM Install - 02 [was: LIBSVM Install - 01 [Was: Re: OCTAVE: LIBSVM and LIBSVM Tools [Win 8 64 bit Pro]]]

[Hans Lodder]

Carnë Draug schreef op 26-12-2013 23:23:
> On 25 Dec 2013 15:35, "Hans Lodder" <address@hidden
> <mailto:address@hidden>> wrote:
>  >
>  > Carnė Draug schreef op 20-12-2013
>  >>
>  >> [...]

Hi Carnë!

I successfully installed and tested LIBSVM 3.17 in OCTAVE 3.6.4. I 
followed the instructions as detailed in 2 LIBSVM readme's, in the 
subdirs windows as well as MATLAB. Reading them very carefully, adding a 
bit of details, and neglecting all the instructions as found on the 
Internet with the help of my best friend Google did the trick. It was 
typically a case of too many and to contradicting information.

I must sincerely apologize to Chih-Chung Chang and Chih-Jen Lin, and to 
you for the noise. Sorry about that.

The PS states how I did it.

I am able to use LIBSVM as I wanted, so my problem is solved. This does 
not change my original offer: If there is anything I can do, please let 
me know.

Kind regards,

Hans Lodder


PS: Install and test LIBSVM

Step 1 is optional. If you have already installed OCTAVE there is no 
need to repeat it. Mind though the directory name.

My platform:

Windows 8 Professional 64-bits
OCTAVE compiled with MingW
Tools: 7-zip, Notepad++

1. Download, install, and test OCTAVE

a. Download octave into the directory C:\octave

Link(MingW):

http://sourceforge.net/projects/octave/files/Octave%20Windows%20binaries/Octave%203.6.4%20for%20Windows%20MinGW%20installer/Octave3.6.4_gcc4.6.2_20130408.7z/download

b. Install OCTAVE

Unzip with 7-zip in place (option 'here')

c. Test OCTAVE

1 + 1

ans = 2

2. Download, install, and test LIBSVM

a. Download LIBSVM into the directory c:\octave\downloads

Link:

http://www.csie.ntu.edu.tw/~cjlin/cgi-bin/libsvm.cgi?+http://www.csie.ntu.edu.tw/~cjlin/libsvm+zip

b. Install LIBSVM

Unzip with 7-zip in place (option 'here')
If not already done so, start OCTAVE. In OCTAVE

ls
cd \octave
cd downloads
cd matlab
make

c. Test LIBSVM

[heart_scale_label, heart_scale_inst] = libsvmread('../heart_scale');  % 
load the test data
model = svmtrain(heart_scale_label, heart_scale_inst, '-c 1 -g 0.07'); % 
create the model
[predict_label, accuracy, dec_values] = svmpredict(heart_scale_label, 
heart_scale_inst, model); % test the training data

Result:
Accuracy = 86.6667% (234/270) (classification)