hex.tree.isofor

Class ModelMetricsAnomaly

java.lang.Object

water.Iced<T>

water.Keyed<hex.ModelMetrics>

hex.ModelMetrics

hex.ModelMetricsUnsupervised

hex.tree.isofor.ModelMetricsAnomaly

All Implemented Interfaces:

hex.ScoreKeeper.ScoreKeeperAware, java.io.Externalizable, java.io.Serializable, java.lang.Cloneable, water.Freezable<hex.ModelMetrics>

public class ModelMetricsAnomaly
extends hex.ModelMetricsUnsupervised
implements hex.ScoreKeeper.ScoreKeeperAware

See Also:

Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	ModelMetricsAnomaly.MetricBuilderAnomaly

Nested classes/interfaces inherited from class hex.ModelMetricsUnsupervised

hex.ModelMetricsUnsupervised.MetricBuilderUnsupervised<T extends hex.ModelMetricsUnsupervised.MetricBuilderUnsupervised<T>>

Nested classes/interfaces inherited from class hex.ModelMetrics

hex.ModelMetrics.MetricBuilder<T extends hex.ModelMetrics.MetricBuilder<T>>

Field Summary

Fields

Modifier and Type	Field and Description
double	_mean_normalized_score Mean normalized score should be (but not necessary is) a number between 0 and 1.
double	_mean_score The raw number that an algorithm is using to count final anomaly score.

Fields inherited from class hex.ModelMetrics

_custom_metric, _description, _MSE, _nobs, _scoring_time

Fields inherited from class water.Keyed

_key

Constructor Summary

Constructors

Constructor and Description
ModelMetricsAnomaly(hex.Model model, water.fvec.Frame frame, hex.CustomMetric customMetric, long nobs, double totalScore, double totalNormScore, java.lang.String description)

Method Summary

Modifier and Type	Method and Description
protected java.lang.StringBuilder	appendToStringMetrics(java.lang.StringBuilder sb)
void	fillTo(hex.ScoreKeeper sk)

Methods inherited from class hex.ModelMetrics

auc_obj, buildKey, buildKey, calcVarImp, calcVarImp, calcVarImp, calcVarImp, checksum_impl, cm, custom_increasing, custom, deepCloneWithDifferentModelAndFrame, defaultModelMetrics, frame, getAllowedMetrics, getFromDKV, getMetricFromModel, getMetricFromModelMetric, hr, isForFrame, isForModel, model, mse, residual_degrees_of_freedom, rmse, setModelKey, sortModelsByMetric, sortModelsByMetric, toString, withDescription, withModelAndFrame

Methods inherited from class water.Keyed

checksum_impl, checksum, checksum, getKey, makeSchema, readAll_impl, readAll, remove_impl, remove_impl, remove_self_key_impl, remove, remove, remove, remove, remove, remove, removeQuietly, writeAll_impl, writeAll

Methods inherited from class water.Iced

asBytes, clone, copyOver, frozenType, read, readExternal, readJSON, reloadFromBytes, toJsonBytes, toJsonString, write, writeExternal, writeJSON

Methods inherited from class java.lang.Object

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

Field Detail

_mean_score

public final double _mean_score

The raw number that an algorithm is using to count final anomaly score. E.g. raw number for Isolation Forest algorithm to count final anomaly score is mean path length of the observation (input row) from root to a leaf.

_mean_normalized_score

public final double _mean_normalized_score

Mean normalized score should be (but not necessary is) a number between 0 and 1. Try to follow convention that higher number means "more anomalous" observation (input row) and number more close to 0 means standard (not anomalous) observation (input row). Always refer to the algorithm's documentation for proper definition of this number. E.g. formula for normalization of Isolation Forest's score is different from the formula for Extended Isolation Forest.

Constructor Detail

ModelMetricsAnomaly

public ModelMetricsAnomaly(hex.Model model,
                           water.fvec.Frame frame,
                           hex.CustomMetric customMetric,
                           long nobs,
                           double totalScore,
                           double totalNormScore,
                           java.lang.String description)

Method Detail

fillTo

public void fillTo(hex.ScoreKeeper sk)

Specified by:

fillTo in interface hex.ScoreKeeper.ScoreKeeperAware

appendToStringMetrics

protected java.lang.StringBuilder appendToStringMetrics(java.lang.StringBuilder sb)

Overrides:

appendToStringMetrics in class hex.ModelMetrics