edu.berkeley.compbio.jlibsvm
Class ImmutableSvmParameter<L extends Comparable,P>

java.lang.Object
  edu.berkeley.compbio.jlibsvm.ImmutableSvmParameter<L,P>

Direct Known Subclasses:

ImmutableSvmParameterGrid, ImmutableSvmParameterPoint

public abstract class ImmutableSvmParameter<L extends Comparable,P>
extends Object

I do not like it

Version:

$Id$

Author:

David Soergel

Nested Class Summary

static class

ImmutableSvmParameter.Builder<L extends Comparable,P>

Field Summary

MultiClassModel.AllVsAllMode	allVsAllMode
float	cache_size
int	crossValidationFolds
float	eps
int	falseClassSVlimit
boolean	gridsearchBinaryMachinesIndependently
double	minVoteProportion
boolean	normalizeL2
float	nu
MultiClassModel.OneVsAllMode	oneVsAllMode
double	oneVsAllThreshold
float	p
boolean	probability
boolean	redistributeUnbalancedC For unbalanced data, redistribute the misclassification cost C according to the numbers of examples in each class, so that each class has the same total misclassification weight assigned to it and the average is param.C
boolean	scaleBinaryMachinesIndependently
ScalingModelLearner<P>	scalingModelLearner When learning scaling, only bother with this many examples, assuming they're in random order.
boolean	shrinking

Constructor Summary

protected

ImmutableSvmParameter(ImmutableSvmParameter.Builder<L,P> copyFrom)

Method Summary

abstract ImmutableSvmParameter.Builder<L,P>	asBuilder()
int	getCacheRows()
Collection<L>	getLabels()
Float	getWeight(L key)
boolean	isWeightsEmpty()
ImmutableSvmParameter<L,P>	noProbabilityCopy()
ImmutableSvmParameter<L,P>	withProbabilityCopy()

Methods inherited from class java.lang.Object

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

Field Detail

cache_size

public final float cache_size

eps

public final float eps

nu

public final float nu

p

public final float p

shrinking

public final boolean shrinking

oneVsAllThreshold

public final double oneVsAllThreshold

oneVsAllMode

public final MultiClassModel.OneVsAllMode oneVsAllMode

allVsAllMode

public final MultiClassModel.AllVsAllMode allVsAllMode

minVoteProportion

public final double minVoteProportion

falseClassSVlimit

public final int falseClassSVlimit

redistributeUnbalancedC

public final boolean redistributeUnbalancedC

For unbalanced data, redistribute the misclassification cost C according to the numbers of examples in each class, so that each class has the same total misclassification weight assigned to it and the average is param.C

scaleBinaryMachinesIndependently

public final boolean scaleBinaryMachinesIndependently

gridsearchBinaryMachinesIndependently

public final boolean gridsearchBinaryMachinesIndependently

normalizeL2

public final boolean normalizeL2

crossValidationFolds

public final int crossValidationFolds

scalingModelLearner

public final ScalingModelLearner<P> scalingModelLearner

When learning scaling, only bother with this many examples, assuming they're in random order.

probability

public final boolean probability

Constructor Detail

ImmutableSvmParameter

protected ImmutableSvmParameter(ImmutableSvmParameter.Builder<L,P> copyFrom)

Method Detail

getCacheRows

public int getCacheRows()

getLabels

public Collection<L> getLabels()

getWeight

public Float getWeight(L key)

isWeightsEmpty

public boolean isWeightsEmpty()

noProbabilityCopy

public ImmutableSvmParameter<L,P> noProbabilityCopy()

withProbabilityCopy

public ImmutableSvmParameter<L,P> withProbabilityCopy()

asBuilder

public abstract ImmutableSvmParameter.Builder<L,P> asBuilder()