org.grouplens.lenskit.eval.metrics.predict

Class NDCGPredictMetric

java.lang.Object

org.grouplens.lenskit.eval.metrics.AbstractMetric<MeanAccumulator,NDCGPredictMetric.Result,NDCGPredictMetric.Result>

org.grouplens.lenskit.eval.metrics.predict.NDCGPredictMetric

All Implemented Interfaces:

Closeable, AutoCloseable, Metric<MeanAccumulator>

public class NDCGPredictMetric
extends AbstractMetric<MeanAccumulator,NDCGPredictMetric.Result,NDCGPredictMetric.Result>

Evaluate a recommender's predictions with normalized discounted cumulative gain.

This is a prediction evaluator that uses base-2 nDCG to evaluate recommender accuracy. The items are ordered by predicted preference and the nDCG is computed using the user's real rating as the gain for each item. Doing this only over the queried items, rather than in the general recommend condition, avoids penalizing recommenders for recommending items that would be better if the user had known about them and provided ratings (e.g., for doing their job).

nDCG is computed per-user and then averaged over all users.

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	NDCGPredictMetric.Result

Constructor Summary

Constructors

Constructor and Description
NDCGPredictMetric()

Method Summary

Modifier and Type	Method and Description
MeanAccumulator	createContext(Attributed algo, TTDataSet ds, Recommender rec) Create the context for an experimental condition (algorithm/data set pair).
NDCGPredictMetric.Result	doMeasureUser(TestUser user, MeanAccumulator context) Measure a user with typed results.
protected NDCGPredictMetric.Result	getTypedResults(MeanAccumulator context) Get the typed results from an accumulator.

Methods inherited from class org.grouplens.lenskit.eval.metrics.AbstractMetric

close, getColumnLabels, getPrefix, getResults, getSuffix, getUserColumnLabels, measureUser

Methods inherited from class java.lang.Object

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

Constructor Detail

NDCGPredictMetric

public NDCGPredictMetric()

Method Detail

createContext

public MeanAccumulator createContext(Attributed algo,
                                     TTDataSet ds,
                                     Recommender rec)

Description copied from interface: Metric

Create the context for an experimental condition (algorithm/data set pair).

Parameters:

algo - The algorithm.

ds - The data set.

rec - The LensKit recommender, if applicable. This can be null for an external algorithm that does not provide a LensKit recommender.

Returns:

The accumulator. This will be passed to Metric.measureUser(TestUser, Object). If the metric does not accumulate any results, this method can return null.

doMeasureUser

public NDCGPredictMetric.Result doMeasureUser(TestUser user,
                                              MeanAccumulator context)

Description copied from class: AbstractMetric

Measure a user with typed results.

Specified by:

doMeasureUser in class AbstractMetric<MeanAccumulator,NDCGPredictMetric.Result,NDCGPredictMetric.Result>

Parameters:

user - The user to measure.

context - The context.

Returns:

The results of measuring the user, or null to emit NAs for the user.

getTypedResults

protected NDCGPredictMetric.Result getTypedResults(MeanAccumulator context)

Description copied from class: AbstractMetric

Get the typed results from an accumulator.

Specified by:

getTypedResults in class AbstractMetric<MeanAccumulator,NDCGPredictMetric.Result,NDCGPredictMetric.Result>

Parameters:

context - The context.

Returns:

The results accumulated for this experiment, or null to emit NAs.