Because you want to use the scripts grid.py and easy.py that come with libsvm, you need to go to official website to download the drawing tool gnuplot, extract it to disk C, enter the directory c:\libsvm\tools, modify the two files grid.py and easy.py with a text editor (Notepad and edit are both acceptable), and if you find the path of gnuplot, modify it according to the actual path and save it.
The connection between python and libsvm (refer to svm's study notes (2) the use of libsvm in python)
1. Open IDLE(pythonGUI) and type >; & gt& gtimportsys & gt& gt& gtsys. Version
2. If your python is 32-bit, the following characters will appear:
(default, apr102012,23: 31:26) [mscv.150032 bit (Intel)]'
At this time, the python interface setting of LIBSVM will be very simple. Find the dynamic link library libsvm.dll in the windows folder under the folder libsvm-3. 16 and add it to the system directory, such as "C:\WINDOWS\system32\", so that you can use libsvm in python.
wk _ ad _ begin({ PID:2 1 })； wk_ad_after(2 1，function(){$('。 ad-hidden)。 hide()； }，function(){$('。 ad-hidden)。 show()； });
Implement a small example
imported goods
Os.chdir ('c: \ libsvm-3.18 \ python') # Please modify it according to the actual path.
fromsvmutilimport *
Y, x = SVM reading problem ('.. /heart _ scale') # Read your own data.
m=svm_train(y[:200]，x[:200]，'-c4 ')
p_label，p_acc，p_val=svm_predict(y[200:]，x[200:]，m)
# # The following results appear, which should be that the installation is correct after optimization, # ITER = 257NU = 0.351161.
obj=-225.628984，rho = 0.636 1 10 nsv = 9 1，nBSV=49
TotalnSV=9 1
Accuracy =84.2857%(59/70) (classification)