org.aksw.limes.core.ml.algorithm.eagle.util

Class CacheTrimmer

java.lang.Object

org.aksw.limes.core.ml.algorithm.eagle.util.CacheTrimmer

public class CacheTrimmer
extends Object

Author:

Klaus Lyko

Constructor Summary

Constructors

Constructor and Description
CacheTrimmer()

Method Summary

Modifier and Type	Method and Description
static AMapping	getRandomTrainingData(AMapping m, int max) Another implementation to get a random training data of size max out of the reference mapping m.
AMapping	getReferenceMapping()
static ACache[]	processData(ACache sC, ACache tC, AMapping m) Method to scale down the Caches used to perform entity matching upon.
ACache[]	processDataEqually(ACache hc1, ACache hc2, AMapping m, int numberOfQuestions)
static AMapping	trimExamples(AMapping m, int max) Method to scale down a reference mapping given by an Oracle.
static AMapping	trimExamplesRandomly(AMapping m, int max) Standard implementation to get random training examples.

Methods inherited from class java.lang.Object

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

Constructor Detail

CacheTrimmer

public CacheTrimmer()

Method Detail

trimExamples

public static AMapping trimExamples(AMapping m,
                                    int max)

Method to scale down a reference mapping given by an Oracle. Only the first max Entries are used.

Parameters:

m - Oracle holding all data.

max - number of examples

Returns:

Mapping scaled down to max entries.

trimExamplesRandomly

public static AMapping trimExamplesRandomly(AMapping m,
                                            int max)

Standard implementation to get random training examples. Basic approach is to get a random set of source uris of the reference mapping and for each one target uri it is mapped to!

Parameters:

m - Mapping

max - number of examples

Returns:

AMapping

getRandomTrainingData

public static AMapping getRandomTrainingData(AMapping m,
                                             int max)

Another implementation to get a random training data of size max out of the reference mapping m. The approach here is to randomly select source URIs of m and for each add ALL target URIs it is mapped to.

Parameters:

m - Mapping

max - number of examples

Returns:

AMapping

processData

public static ACache[] processData(ACache sC,
                                   ACache tC,
                                   AMapping m)

Method to scale down the Caches used to perform entity matching upon. Scaling down is done according to the given reference mapping. Returns Caches only holding instances of the reference Mapping.

Parameters:

sC - Cache for source data.

tC - Cache for target data.

m - Reference Mapping (e.g. part of the optimal mapping)

Returns:

Array holding both resulting Caches, where the Cache for the source is at index 0. Cache for the target knowledge base at index 1.

processDataEqually

public ACache[] processDataEqually(ACache hc1,
                                   ACache hc2,
                                   AMapping m,
                                   int numberOfQuestions)

getReferenceMapping

public AMapping getReferenceMapping()

Returns:

reference mapping