Package pt.tumba.ngram.svm

Implementation of Support Vector Machines classification and regression that can be used to categorize text files using N-Grams as features.

See:
Description

Class Summary

Cache	SVM Kernel Cache, implementing a least recently used (LRU) policy.
Kernel	Abstract interface to model an SVM Kernel.
ONECLASSQ	Inner class representing a Kernel matrix for SVM Distribution Estimation.
Solver	Generalized SMO+SVMlight algorithm Solves: min 0.5(\alpha^T Q \alpha) + b^T \alpha y^T \alpha = \delta y_i = +1 or -1 0 <= alpha_i <= Cp for y_i = 1 0 <= alpha_i <= Cn for y_i = -1 Given: Q, b, y, Cp, Cn, and an initial feasible point \alpha l is the size of vectors and matrices eps is the stopping criterion solution will be put in \alpha, objective value will be put in obj
Solver.SolutionInfo
SolverNU	Solver for nu-svm classification and regression additional constraint: e^T \alpha = constant
SVCQ	Inner class representing a Kernel matrix for Support Vector classification.
SVM	Construct and solve various formulations of the support vector machine (SVM) problem.
SVM.decisionFunction	Inner class modeling the data for the SVM decision function, used for classifying points with respect to the hyperplane.
SVMCategorizer	Simple, easy-to-use, and efficient software for SVM classification and regression, based on the LIBSVM implementation of Chin-Chung Chang and Chin-Jen Lin.
SVMModel	SVMModel encondes a classification model, describing both the model parameters and the Support Vectors.
SVMNode	SVMNode is used to model dimentions in vectors.
SVMParameter	Constants and Parameters used used in the SVM package.
SVMProblem	Class to model an SVM Problem, containing both the training vectors and the class (value) associated with each vector.
SVRQ	Inner class representing a Kernel matrix for Support Vector regression.

Package pt.tumba.ngram.svm Description

Implementation of Support Vector Machines classification and regression that can be used to categorize text files using N-Grams as features.
A support vector machine is a supervised learning algorithm developed over the past decade by Vapnik and others (Vapnik, Statistical Learning Theory, 1998). It operates by mapping the given training set into a possibly high-dimensional feature space and attempting to locate in that space a plane that separates the positive from the negative examples.