ScoreKeeper (h2o-core 3.42.0 API)

java.lang.Object
- water.Iced
- - hex.ScoreKeeper

All Implemented Interfaces:

java.io.Externalizable, java.io.Serializable, java.lang.Cloneable, Freezable
```
public class ScoreKeeper
extends Iced
```
Low-weight keeper of scores solely intended for display (either direct or as helper to create scoring history TwoDimTable). Not intended to store large AUC object or ConfusionMatrices, etc.

See Also:

Serialized Form

Nested Class Summary

Nested Classes
Modifier and Type	Class and Description
`static interface`	`ScoreKeeper.IStoppingMetric`
`static class`	`ScoreKeeper.ProblemType`
`static interface`	`ScoreKeeper.ScoreKeeperAware`
`static class`	`ScoreKeeper.StoppingMetric`

Field Summary

Fields
Modifier and Type	Field and Description
`double`	`_anomaly_score`
`double`	`_anomaly_score_normalized`
`double`	`_AUC`
`double`	`_AUUC`
`int`	`_auuc_nbins`
`double`	`_auuc_normalized`
`double`	`_classError`
`double`	`_custom_metric`
`float[]`	`_hitratio`
`double`	`_lift`
`double`	`_logloss`
`double`	`_mae`
`double`	`_mean_per_class_error`
`double`	`_mean_residual_deviance`
`double`	`_mse`
`double`	`_pr_auc`
`double`	`_qini`
`double`	`_r2`
`double`	`_rmse`
`double`	`_rmsle`

Constructor Summary

Constructors
Constructor and Description
`ScoreKeeper()`
`ScoreKeeper(double mse)` Keep score of mean squared error only.
`ScoreKeeper(Model m)` Keep score for a model using its validation_metrics if available and training_metrics if not.
`ScoreKeeper(ModelMetrics mm)` Keep score of a given ModelMetrics.

Method Summary

All Methods Static Methods Instance Methods Concrete Methods
Modifier and Type	Method and Description
`static int`	`best(ScoreKeeper[] sk, int k, ScoreKeeper.IStoppingMetric criterion)` Based on the given array of ScoreKeeper and stopping criteria what is the best scoring iteration of the last k iterations?
`boolean`	`equals(java.lang.Object that)` Compare this ScoreKeeper with that ScoreKeeper
`void`	`fillFrom(ModelMetrics m)`
`void`	`fillFrom(ModelMetrics m, CustomMetric customMetric)`
`boolean`	`isEmpty()`
`static boolean`	`stopEarly(ScoreKeeper[] sk, int k, ScoreKeeper.ProblemType type, ScoreKeeper.IStoppingMetric criterion, double rel_improvement, java.lang.String what, boolean verbose)` Based on the given array of ScoreKeeper and stopping criteria should we stop early?
`java.lang.String`	`toString()`

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

Field Detail

_mean_residual_deviance
```
public double _mean_residual_deviance
```

_mse
```
public double _mse
```

_rmse
```
public double _rmse
```

_mae
```
public double _mae
```

_rmsle
```
public double _rmsle
```

_logloss
```
public double _logloss
```

_AUC
```
public double _AUC
```

_pr_auc
```
public double _pr_auc
```

_classError
```
public double _classError
```

_mean_per_class_error
```
public double _mean_per_class_error
```

_custom_metric
```
public double _custom_metric
```

_hitratio
```
public float[] _hitratio
```

_lift
```
public double _lift
```

_r2
```
public double _r2
```

_anomaly_score
```
public double _anomaly_score
```

_anomaly_score_normalized

public double _anomaly_score_normalized

_AUUC
```
public double _AUUC
```

_auuc_normalized
```
public double _auuc_normalized
```

_qini
```
public double _qini
```

_auuc_nbins
```
public int _auuc_nbins
```

Constructor Detail
- ScoreKeeper
```
public ScoreKeeper()
```
- ScoreKeeper
```
public ScoreKeeper(double mse)
```
  Keep score of mean squared error only.
  
  Parameters:
  
  mse -
- ScoreKeeper
```
public ScoreKeeper(ModelMetrics mm)
```
  Keep score of a given ModelMetrics.
  
  Parameters:
  
  mm - ModelMetrics to keep track of.
- ScoreKeeper
```
public ScoreKeeper(Model m)
```
  Keep score for a model using its validation_metrics if available and training_metrics if not.
  
  Parameters:
  
  m - model for which we should keep score

Method Detail

isEmpty
```
public boolean isEmpty()
```

fillFrom
```
public void fillFrom(ModelMetrics m)
```

fillFrom

public void fillFrom(ModelMetrics m,
                     CustomMetric customMetric)

best

public static int best(ScoreKeeper[] sk,
                       int k,
                       ScoreKeeper.IStoppingMetric criterion)

Based on the given array of ScoreKeeper and stopping criteria what is the best scoring iteration of the last k iterations?

stopEarly

public static boolean stopEarly(ScoreKeeper[] sk,
                                int k,
                                ScoreKeeper.ProblemType type,
                                ScoreKeeper.IStoppingMetric criterion,
                                double rel_improvement,
                                java.lang.String what,
                                boolean verbose)

Based on the given array of ScoreKeeper and stopping criteria should we stop early?

equals
```
public boolean equals(java.lang.Object that)
```
Compare this ScoreKeeper with that ScoreKeeper

Overrides:

equals in class java.lang.Object

Parameters:

that -

Returns:

true if they are equal (up to 1e-6 absolute and relative error, or both contain NaN for the same values)

toString
```
public java.lang.String toString()
```
Overrides:

toString in class java.lang.Object

Class ScoreKeeper

Nested Class Summary

Field Summary

Constructor Summary

Method Summary

Methods inherited from class water.Iced

Methods inherited from class java.lang.Object

Field Detail

_mean_residual_deviance

_mse

_rmse

_mae

_rmsle

_logloss

_AUC

_pr_auc

_classError

_mean_per_class_error

_custom_metric

_hitratio

_lift

_r2

_anomaly_score

_anomaly_score_normalized

_AUUC

_auuc_normalized

_qini

_auuc_nbins

Constructor Detail

ScoreKeeper

ScoreKeeper

ScoreKeeper

ScoreKeeper

Method Detail

isEmpty

fillFrom

fillFrom

best

stopEarly

equals

toString