org.lenskit.eval.traintest.recommend

Class ListOnlyTopNMetric<X>

java.lang.Object

org.lenskit.eval.traintest.metrics.Metric<X>

org.lenskit.eval.traintest.recommend.TopNMetric<X>

org.lenskit.eval.traintest.recommend.ListOnlyTopNMetric<X>

Direct Known Subclasses:

TopNEntropyMetric, TopNLengthMetric, TopNMAPMetric, TopNMRRMetric, TopNNDCGMetric, TopNNDPMMetric, TopNPopularityMetric, TopNPrecisionRecallMetric

public abstract class ListOnlyTopNMetric<X>
extends TopNMetric<X>

Intermediate class for top-N metrics that only depend on the list of recommended items, not their details. Metrics extending this class will implement the measureUser(Recommender, TestUser, int, LongList, Object) method instead of measureUser(Recommender, TestUser, int, ResultList, Object). The recommend eval task uses this subclass to improve efficiency when results are not used in the evaluation.

Constructor Summary

Constructors

Modifier	Constructor and Description
protected	ListOnlyTopNMetric(java.lang.Class<? extends TypedMetricResult> resType, java.lang.Class<? extends TypedMetricResult> aggType)
protected	ListOnlyTopNMetric(java.lang.Class<? extends TypedMetricResult> resType, java.lang.Class<? extends TypedMetricResult> aggType, java.lang.String suffix)
protected	ListOnlyTopNMetric(java.util.List<java.lang.String> labels, java.util.List<java.lang.String> aggLabels)

Method Summary

Modifier and Type	Method and Description
abstract MetricResult	measureUser(Recommender rec, TestUser user, int targetLength, LongList recommendations, X context) Measurement method that only uses the recommend list.
MetricResult	measureUser(Recommender rec, TestUser user, int targetLength, ResultList recommendations, X context) Measure a single result.

Methods inherited from class org.lenskit.eval.traintest.metrics.Metric

createContext, getAggregateColumnLabels, getAggregateMeasurements, getColumnLabels, getRequiredRoots

Methods inherited from class java.lang.Object

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

Constructor Detail

ListOnlyTopNMetric

protected ListOnlyTopNMetric(java.util.List<java.lang.String> labels,
                             java.util.List<java.lang.String> aggLabels)

ListOnlyTopNMetric

protected ListOnlyTopNMetric(java.lang.Class<? extends TypedMetricResult> resType,
                             java.lang.Class<? extends TypedMetricResult> aggType)

ListOnlyTopNMetric

protected ListOnlyTopNMetric(java.lang.Class<? extends TypedMetricResult> resType,
                             java.lang.Class<? extends TypedMetricResult> aggType,
                             java.lang.String suffix)

Method Detail

measureUser

@Nonnull
public final MetricResult measureUser(Recommender rec,
                                               TestUser user,
                                               int targetLength,
                                               ResultList recommendations,
                                               X context)

Description copied from class: TopNMetric

Measure a single result. The result may come from either prediction or recommendation.

Thread Safety: This method may be called concurrently by multiple threads with the same recommender and context.

Specified by:

measureUser in class TopNMetric<X>

Parameters:

rec - The recommender used to recommend for this user.

user - The user’s test data.

targetLength - The intended length of the recommendation list.

recommendations - The user’s recommendations.

Returns:

A list of fields to add to the result’s output.

measureUser

@Nonnull
public abstract MetricResult measureUser(Recommender rec,
                                                  TestUser user,
                                                  int targetLength,
                                                  LongList recommendations,
                                                  X context)

Measurement method that only uses the recommend list.

Thread Safety: This method may be called concurrently by multiple threads with the same recommender and context.

Parameters:

rec - The recommender used to recommend for this user.

user - The user.

targetLength - The target list length.

recommendations - The list of recommendations.

context - The context.

Returns:

The results of measuring this user.