Package it.unimi.di.big.mg4j.search.score

Class LinearAggregator

java.lang.Object

it.unimi.di.big.mg4j.search.score.AbstractAggregator

it.unimi.di.big.mg4j.search.score.LinearAggregator

All Implemented Interfaces:

Scorer, FlyweightPrototype<Scorer>

public class LinearAggregator
extends AbstractAggregator

An aggregator that computes a linear combination of the component scorers.

This class requires, beside the usually array of scorers, a parallel array of weights (not to be confused with index weights). The score from each scorer will be multiplied by the respective weight, and the overal score will be the sum of these values. Equalisation, if necessary, if performed by maximising over the sample scores and dividing all scores by the resulting value.

Field Summary

Fields

Modifier and Type	Field	Description
protected double[]	equalisationFactor	The equalisation factors for each scorer (all set to one if no equalisation is required).
protected double[]	weight	The weights of each scorer.

Fields inherited from class it.unimi.di.big.mg4j.search.score.AbstractAggregator

actualSamples, currSample, currScore, documentIterator, n, needsCaching, sampleDocument, samples, sampleScore, scorer

Constructor Summary

Constructors

Constructor	Description
LinearAggregator(Scorer[] scorer, double[] weight)	Creates a linear aggregator.

Method Summary

Modifier and Type	Method	Description
LinearAggregator	copy()
protected double	score(double[] score)	Computes an aggregated score using the given array of basic scores.
protected void	setupEqualizationFactors()	Sets up the equalisation factors.
String	toString()

Methods inherited from class it.unimi.di.big.mg4j.search.score.AbstractAggregator

equalize, getWeights, nextDocument, score, score, setWeights, usesIntervals, wrap

Methods inherited from class java.lang.Object

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

Field Detail

weight

protected final double[] weight

The weights of each scorer.

equalisationFactor

protected final double[] equalisationFactor

The equalisation factors for each scorer (all set to one if no equalisation is required).

Constructor Detail

LinearAggregator

public LinearAggregator(Scorer[] scorer,
                        double[] weight)

Creates a linear aggregator.

Parameters:

scorer - the array of scorers.

weight - the array of weights.

Method Detail

copy

public LinearAggregator copy()

score

protected double score(double[] score)

Description copied from class: AbstractAggregator

Computes an aggregated score using the given array of basic scores. The array is parallel to AbstractAggregator.scorer.

Specified by:

score in class AbstractAggregator

Parameters:

score - an array of scores.

Returns:

the aggregated scorer.

setupEqualizationFactors

protected void setupEqualizationFactors()

Description copied from class: AbstractAggregator

Sets up the equalisation factors.

Implementations should look into AbstractAggregator.sampleScore and set up the equalisation logic. Note that this method is responsible for setting up appropriate equalisation factors even if no equalisation is required (e.g., setting all factors to 1 ).

Specified by:

setupEqualizationFactors in class AbstractAggregator

toString

public String toString()

Overrides:

toString in class Object