org.tribuo.anomaly.evaluation

Class AnomalyMetric

java.lang.Object

org.tribuo.anomaly.evaluation.AnomalyMetric

All Implemented Interfaces:

EvaluationMetric<Event,org.tribuo.anomaly.evaluation.AnomalyMetric.Context>

public class AnomalyMetric
extends Object
implements EvaluationMetric<Event,org.tribuo.anomaly.evaluation.AnomalyMetric.Context>

A metric for evaluating anomaly detection problems. The sufficient statistics must be encoded in the number of true positives, false positives, true negatives and false negatives.

Nested Class Summary

Nested classes/interfaces inherited from interface org.tribuo.evaluation.metrics.EvaluationMetric

EvaluationMetric.Average

Constructor Summary

Constructors

Constructor and Description
AnomalyMetric(MetricTarget<Event> target, String name, ToDoubleBiFunction<MetricTarget<Event>,org.tribuo.anomaly.evaluation.AnomalyMetric.Context> impl) Creates an anomaly detection metric, with a specific name, using the supplied evaluation function.

Method Summary

Modifier and Type	Method and Description
double	compute(org.tribuo.anomaly.evaluation.AnomalyMetric.Context context) Compute the result of this metric from the input context.
org.tribuo.anomaly.evaluation.AnomalyMetric.Context	createContext(Model<Event> model, List<Prediction<Event>> predictions) Creates the context this metric uses to compute it's value.
String	getName() The name of this metric.
MetricTarget<Event>	getTarget() The target for this metric instance.

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.tribuo.evaluation.metrics.EvaluationMetric

createContext, getID

Constructor Detail

AnomalyMetric

public AnomalyMetric(MetricTarget<Event> target,
                     String name,
                     ToDoubleBiFunction<MetricTarget<Event>,org.tribuo.anomaly.evaluation.AnomalyMetric.Context> impl)

Creates an anomaly detection metric, with a specific name, using the supplied evaluation function.

Parameters:

target - The target of the metric (i.e., the event type or an average).

name - The name of the metric.

impl - The implementation function.

Method Detail

compute

public double compute(org.tribuo.anomaly.evaluation.AnomalyMetric.Context context)

Description copied from interface: EvaluationMetric

Compute the result of this metric from the input context.

Specified by:

compute in interface EvaluationMetric<Event,org.tribuo.anomaly.evaluation.AnomalyMetric.Context>

Parameters:

context - The context to use.

Returns:

The value of the metric.

getTarget

public MetricTarget<Event> getTarget()

Description copied from interface: EvaluationMetric

The target for this metric instance.

Specified by:

getTarget in interface EvaluationMetric<Event,org.tribuo.anomaly.evaluation.AnomalyMetric.Context>

Returns:

The metric target.

getName

public String getName()

Description copied from interface: EvaluationMetric

The name of this metric.

Specified by:

getName in interface EvaluationMetric<Event,org.tribuo.anomaly.evaluation.AnomalyMetric.Context>

Returns:

The name.

createContext

public org.tribuo.anomaly.evaluation.AnomalyMetric.Context createContext(Model<Event> model,
                                                                         List<Prediction<Event>> predictions)

Description copied from interface: EvaluationMetric

Creates the context this metric uses to compute it's value.

Specified by:

createContext in interface EvaluationMetric<Event,org.tribuo.anomaly.evaluation.AnomalyMetric.Context>

Parameters:

model - The model to use.

predictions - The predictions to use.

Returns:

The metric context.