ai.libs.jaicore.ml.dyadranking.optimizing

Class BilinFunction

java.lang.Object

ai.libs.jaicore.ml.dyadranking.optimizing.BilinFunction

All Implemented Interfaces:

edu.stanford.nlp.optimization.DiffFunction, edu.stanford.nlp.optimization.Function

public class BilinFunction
extends java.lang.Object
implements edu.stanford.nlp.optimization.DiffFunction

Wraps the NLL optimizing problem into the QNMinimizer optimizer.

Constructor Summary

Constructors

Constructor and Description
BilinFunction(java.util.Map<IDyadRankingInstance,java.util.Map<Dyad,ai.libs.jaicore.math.linearalgebra.Vector>> featureTransforms, DyadRankingDataset drDataset, int dimension) Creates a NLL optimizing problem for the kronecker product as the bilinear feature transform.

Method Summary

Modifier and Type	Method and Description
double[]	derivativeAt(double[] x)
int	domainDimension()
double	valueAt(double[] x)

Methods inherited from class java.lang.Object

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

Constructor Detail

BilinFunction

public BilinFunction(java.util.Map<IDyadRankingInstance,java.util.Map<Dyad,ai.libs.jaicore.math.linearalgebra.Vector>> featureTransforms,
                     DyadRankingDataset drDataset,
                     int dimension)

Creates a NLL optimizing problem for the kronecker product as the bilinear feature transform.

Parameters:

featureTransform - the feature transform, must be an instance of BiliniearFeatureTransform

drDataset - the dataset to optimize

dimension - the dimension of the optimized vector

Method Detail

valueAt

public double valueAt(double[] x)

Specified by:

valueAt in interface edu.stanford.nlp.optimization.Function

domainDimension

public int domainDimension()

Specified by:

domainDimension in interface edu.stanford.nlp.optimization.Function

derivativeAt

public double[] derivativeAt(double[] x)

Specified by:

derivativeAt in interface edu.stanford.nlp.optimization.DiffFunction