SVM Types

分类 C_SVC=100: C-Support Vector Classification. n-class classification (n $\ge$ 2), allows imperfect separation of classes with penalty multiplier C for outliers.

NU_SVC=101: $\nu$-Support Vector Classification. n-class classification with possible imperfect separation. Parameter $\nu$ (in the range 0…1, the larger the value, the smoother the decision boundary) is used instead of C.

单类界限 ONE_CLASS=102: Distribution Estimation (One-class %SVM). All the training data are from the same class, %SVM builds a boundary that separates the class from the rest of the feature space.

回归 EPS_SVR=103: $ϵ$-Support Vector Regression. The distance between feature vectors from the training set and the fitting hyper-plane must be less than p. For outliers the penalty multiplier C is used.

NU_SVR=104: $\nu$-Support Vector Regression. $\nu$ is used instead of p.

SVM KernelType

CUSTOM=-1: Returned by SVM::getKernelType in case when custom kernel has been set

LINEAR=0: Linear kernel. No mapping is done, linear discrimination (or regression) is done in the original feature space. It is the fastest option. $K(x_i,x_j)=x_i^T{x}_j$.

POLY=1: Polynomial kernel: $K(x_i,x_j)=(\gamma x_i^T{x}_j+coef0{)}^{degree},\gamma >0$.

RBF=2: Radial basis function (RBF), a good choice in most cases.$K(x_i,x_j)=e^{-\gamma ||x_i-x_j|{|}^2},\gamma >0$.

SIGMOID=3: Sigmoid kernel: $K(x_i,x_j)=\tanh (\gamma x_i^T{x}_j+coef0)$.

CHI2=4: Exponential Chi2 kernel, similar to the RBF kernel: $K(x_i,x_j)=e^{-\gamma {\chi }^2(x_i,x_j)},{\chi }^2(x_i,x_j)=(x_i-x_j{)}^2/(x_i+x_j),\gamma >0$.

INTER=5: Histogram intersection kernel. A fast kernel. $K(x_i,x_j)=min(x_i,x_j)$.

SVM ParamTypes

KernelType参数：

GAMMA=1,COEF=4,DEGREE=5

SVM Types参数：

C=0: 取值为$(0,+\infty )$？,数值越大，软间隔越小P=2, 取值为$(0,1)$？,数值越大，软间隔越大NU=3: 取值为$(0,1)$,数值越大，软间隔越大