MaltParser 1.5.1

org.maltparser.ml.lib
Class MaltLibsvmModel

java.lang.Object
  extended by org.maltparser.ml.lib.MaltLibsvmModel
All Implemented Interfaces:
Serializable, MaltLibModel

public class MaltLibsvmModel
extends Object
implements Serializable, MaltLibModel

This class borrows code from libsvm.svm.java of the Java implementation of the libsvm package. MaltLibsvmModel stores the model obtained from the training procedure. In addition to the original code the model is more integrated to MaltParser. Instead of moving features from MaltParser's internal data structures to liblinear's data structure it uses MaltParser's data structure directly on the model.

Author:
Johan Hall
See Also:
Serialized Form

Field Summary
 int l
           
 int[] label
           
 int nr_class
           
 int[] nSV
           
 libsvm.svm_parameter param
           
 double[] rho
           
 int[] start
           
 libsvm.svm_node[][] SV
           
 double[][] sv_coef
           
 
Constructor Summary
MaltLibsvmModel(libsvm.svm_model model, libsvm.svm_problem problem)
           
 
Method Summary
 int[] getLabels()
           
 int[] predict(MaltFeatureNode[] x)
           
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Field Detail

param

public libsvm.svm_parameter param

nr_class

public int nr_class

l

public int l

SV

public libsvm.svm_node[][] SV

sv_coef

public double[][] sv_coef

rho

public double[] rho

label

public int[] label

nSV

public int[] nSV

start

public int[] start
Constructor Detail

MaltLibsvmModel

public MaltLibsvmModel(libsvm.svm_model model,
                       libsvm.svm_problem problem)
Method Detail

predict

public int[] predict(MaltFeatureNode[] x)
Specified by:
predict in interface MaltLibModel

getLabels

public int[] getLabels()

MaltParser 1.5.1

Copyright 2007-2011 Johan Hall, Jens Nilsson and Joakim Nivre.