1. Zhenshan, Wen wenzs@pku.edu.cn
SVM Implementation
Zhenshan, Wen

2. Support Vector Machine (SVM)
Name: Support Vector Machine (SVM)
Links: SVMs at SVMs at Bell Labs SVMs at Microsoft Research SVMs at Royal Holloway College Learning Systms Group ANU SVMs at the MIT SVMs at the Bristol CI-Group
Publications:
Burges, C. (1998) A Tutorial on Support Vector Machines for Pattern Recognition
Vapnik, Vladimir (1998): Statistical Learning Theory
Systems: SVM light, mySVM
HypothesisLanguage: Functions
Tasks: Concept Learning ,Function Approximation

3. LibSVM LIBSVM is an integrated software Support vector classification
C-SVC
nu-SVC
Support vector regression
one-class SVM
epsilon-SVR
nu-SVR
Also supports multi-class classification
Basic algorithm is a simplification stemmed from
SMO, by Platt
SVMLight, by Joachims
modification 2 of SMO, by Keerthi et al

4. LibSVM Main features of LIBSVM include : Different SVM formulations
Efficient multi-class classification
Cross validation for model selection
Weighted SVM for unbalanced data
Both C++ and Java sources
GUI demonstrating SVM classification and regression
Python, R (also Splus), Matlab, Perl, and Ruby interfaces
Automatic model selection which can generate contour of cross valiation accuracy

6. svm_train Usage
Usage: svm_train [options] training_set_file [model_file], options:
-s svm_type : set type of SVM (default 0)
0 -- C-SVC
1 -- nu-SVC
2 -- one-class SVM
3 -- epsilon-SVR
4 -- nu-SVR
-t kernel_type : set type of kernel function (default 2)
0 -- linear: u'*v
1 -- polynomial: (gamma*u'*v + coef0)^degree
2 -- radial basis function: exp(-gamma*|u-v|^2)
3 -- sigmoid: tanh(gamma*u'*v + coef0)
-d degree : set degree in kernel function (default 3)
-g gamma : set gamma in kernel function (default 1/k)
-r coef0 : set coef0 in kernel function (default 0)
-c cost : set the parameter C of C-SVC, epsilon-SVR, and nu-SVR (default 1)
-n nu : set the parameter nu of nu-SVC, one-class SVM, and nu-SVR (default 0.5)
-p epsilon : set the epsilon in loss function of epsilon-SVR (default 0.1)
-m cachesize : set cache memory size in MB (default 40)
-e epsilon : set tolerance of termination criterion (default 0.001)
-h shrinking: whether to use the shrinking heuristics, 0 or 1 (default 1)
-wi weight: set the parameter C of class i to weight*C in C-SVC (default 1)
-v n: n-fold cross validation mode

7. svm_predict Usage Usage: svm-predict test_file model_file output_file
model_file is the model file generated by svm-train.
test_file is the test data you want to predict.
svm-predict will produce output in the output_file.
Tips on practical use
=====================
* Scale your data. For example, scale each attribute to [0,1] or [-1,+1].
* For C-SVC, consider using the model selection tool in the python directory.
* nu in nu-SVC/one-class-SVM/nu-SVR approximates the fraction of training
errors and support vectors.
* If data for classification are unbalanced (e.g. many positive and
few negative), try different penalty parameters C by -wi (see
examples below).

8. Example (c version)
> svm-train -s 0 -c t 1 -g 1 -r 1 -d 3 data_file
Train a classifier with polynomial kernel (u'v+1)^3 and C = 1000
> svm-train -s 1 -n 0.1 -t 2 -g 0.5 -e data_file
Train a classifier by nu-SVM (nu = 0.1) with RBF kernel exp(-0.5|u-v|^2) and stopping tolerance
> svm-train -s 3 -p 0.1 -t 0 -c 10 data_file
Solve SVM regression with linear kernel u'v and C=10, and epsilon = 0.1 in the loss function.
> svm-train -s 0 -c 10 -w1 1 -w-1 5 data_file
Train a classifier with penalty 10 for class 1 and penalty 50 for class -1.
> svm-train -s 0 -c 500 -g 0.1 -v 5 data_file
Do five-fold cross validation for the classifier using the parameters C = 500 and gamma = 0.1

9. svm Java Class Class svm provides some important methods: svm_train
svm_predict
svm_save_model
svm_load_model
svm_check_parameter

10. Data Representation
Using the same representation of training data as SVM_light uses.
BNF-like representation
<class> .=. +1 | -1
<feature> .=. integer (>=1)
<value> .=. real
<line> .=. <class> <feature>:<value><feature>:<value> ... <feature>:<value>

11. Data Representation Example
Example (SVM)
+1 201: : :1 4882:1
-1 874: : :1 6179:1
Example (SVR)
: : :1 4882:1
: : :1 6179:1