org.lenskit.eval.traintest.predict

Class PredictMetric<X>

java.lang.Object

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

org.lenskit.eval.traintest.predict.PredictMetric<X>

Direct Known Subclasses:

CoveragePredictMetric, MAEPredictMetric, NDCGPredictMetric, RMSEPredictMetric

public abstract class PredictMetric<X>
extends Metric<X>

Base class for metrics that measure predictions for users.

Constructor Summary

Constructors

Modifier	Constructor and Description
protected	PredictMetric(Class<? extends TypedMetricResult> resType, Class<? extends TypedMetricResult> aggType) Construct a new result metric.
protected	PredictMetric(List<String> labels, List<String> aggLabels) Construct a new result metric.

Method Summary

Modifier and Type	Method and Description
abstract MetricResult	measureUser(TestUser user, ResultMap predictions, 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

PredictMetric

protected PredictMetric(List<String> labels,
                        List<String> aggLabels)

Construct a new result metric.

Parameters:

labels - Column labels.

aggLabels - Aggregate column labels.

PredictMetric

protected PredictMetric(Class<? extends TypedMetricResult> resType,
                        Class<? extends TypedMetricResult> aggType)

Construct a new result metric.

Parameters:

resType - The result type for measuring results, or null for no measurement.

aggType - The result type for aggregate measurements, or null for no measurement.

Method Detail

measureUser

@Nonnull
public abstract MetricResult measureUser(TestUser user,
                                                   ResultMap predictions,
                                                   X context)

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

Parameters:

user - The user’s test data.

predictions - The predictions.

Returns:

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