(PECL svm >= 0.1.0)
SVM::C_SVCThe basic C_SVC SVM type. The default, and a good starting point
SVM::NU_SVCThe NU_SVC type uses a different, more flexible, error weighting
SVM::ONE_CLASSOne class SVM type. Train just on a single class, using outliers as negative examples
SVM::EPSILON_SVRA SVM type for regression (predicting a value rather than just a class)
SVM::NU_SVRA NU style SVM regression type
SVM::KERNEL_LINEARA very simple kernel, can work well on large document classification problems
SVM::KERNEL_POLYA polynomial kernel
SVM::KERNEL_RBFThe common Gaussian RBD kernel. Handles non-linear problems well and is a good default for classification
SVM::KERNEL_SIGMOIDA kernel based on the sigmoid function. Using this makes the SVM very similar to a two layer sigmoid based neural network
SVM::KERNEL_PRECOMPUTEDA precomputed kernel - currently unsupported.
SVM::OPT_TYPEThe options key for the SVM type
SVM::OPT_KERNEL_TYPEThe options key for the kernel type
SVM::OPT_DEGREESVM::OPT_SHRINKINGTraining parameter, boolean, for whether to use the shrinking heuristics
SVM::OPT_PROBABILITYTraining parameter, boolean, for whether to collect and use probability estimates
SVM::OPT_GAMMAAlgorithm parameter for Poly, RBF and Sigmoid kernel types.
SVM::OPT_NUThe option key for the nu parameter, only used in the NU_ SVM types
SVM::OPT_EPSThe option key for the Epsilon parameter, used in epsilon regression
SVM::OPT_PTraining parameter used by Episilon SVR regression
SVM::OPT_COEF_ZEROAlgorithm parameter for poly and sigmoid kernels
SVM::OPT_CThe option for the cost parameter that controls tradeoff between errors and generality - effectively the penalty for misclassifying training examples.
SVM::OPT_CACHE_SIZEMemory cache size, in MB