se.lth.cs.srl.ml.liblinear

Class WeightVector

java.lang.Object

se.lth.cs.srl.ml.liblinear.WeightVector

All Implemented Interfaces:

Serializable

Direct Known Subclasses:

WeightVector.BinaryVector, WeightVector.MultipleVector

public abstract class WeightVector
extends Object
implements Serializable

See Also:

Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	WeightVector.BinaryLibLinearVector
static class	WeightVector.BinarySparseVector
static class	WeightVector.BinaryVector
static class	WeightVector.MultipleLibLinearVector
static class	WeightVector.MultipleSparseVector
static class	WeightVector.MultipleVector

Field Summary

Fields

Modifier and Type	Field and Description
protected double	bias
protected int	classes
protected int	features

Constructor Summary

Constructors

Constructor and Description
WeightVector(double bias, int features, int classes)

Method Summary

Modifier and Type	Method and Description
abstract double[]	computeAllProbs(Collection<Integer> ints)
abstract short	computeBestClass(Collection<Integer> ints)
static WeightVector	parseWeights(BufferedReader in, int features, int classes, double bias, boolean sparse) Used to parse the weights and return the proper weightvector

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

bias

protected double bias

features

protected int features

classes

protected int classes

Constructor Detail

WeightVector

public WeightVector(double bias,
                    int features,
                    int classes)

Method Detail

parseWeights

public static WeightVector parseWeights(BufferedReader in,
                                        int features,
                                        int classes,
                                        double bias,
                                        boolean sparse)
                                 throws IOException

Used to parse the weights and return the proper weightvector

Parameters:

in - the inputstream, assuming header is parsed and weights are next

features - number of features

classes - number of classes

w - the w-string (last line before the weights), should be either 'w' or 'w-sparse'

Returns:

the appropriate weightvector

Throws:

IOException

computeAllProbs

public abstract double[] computeAllProbs(Collection<Integer> ints)

computeBestClass

public abstract short computeBestClass(Collection<Integer> ints)