Using LIBSVM under Linux

1. sudo apt-get install Python-pip

2. sudo apt-get install Python-dev

3. sudo apt-get install gnuplot

4. Download libsvm-3.20 Extract, add this extract folder CP to/usr/bin

5. Which gnuplot view the path result for this command is/usr/bin/gnuplot

Then look at the two files under Libsvm-3.20/python easy.py and grid.py about gnuplot in the path Gnuplot_exe = "/usr/bin/gnuplot" and just which out of the same result, so no need to modify

6. Go to/user/bin/libsvm-3.20 to execute make

7. Go to/user/bin/libsvm-3.20/python to execute make

8. Download the installation apt-get install Libsvm-tools

9. Enter the/usr/bin/libsvm-3.20/tools directory

Parameters of the tuning c,g

Python grid.py file name

C For example, 0.0005 for the G.

10. Derive the SVM model

Svm-train-c 5000-g 0.0005 Training set file name Train.model

11. Testing the SVM model

Svm-predict Test Set file name Train.model test.predict

Using LIBSVM under Linux