org.tribuo.sequence

Class SequenceModel<T extends Output<T>>

java.lang.Object

org.tribuo.sequence.SequenceModel<T>

Type Parameters:

T - the type of the outputs used to train the model. SequenceModel implementations must be serializable!

All Implemented Interfaces:

com.oracle.labs.mlrg.olcut.provenance.Provenancable<ModelProvenance>, Serializable

Direct Known Subclasses:

ConfidencePredictingSequenceModel, TensorflowSequenceModel, ViterbiModel

public abstract class SequenceModel<T extends Output<T>>
extends Object
implements com.oracle.labs.mlrg.olcut.provenance.Provenancable<ModelProvenance>, Serializable

A prediction model, which is used to predict outputs for unseen instances.

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
protected ImmutableFeatureMap	featureIDMap
protected String	name
protected ImmutableOutputInfo<T>	outputIDMap
protected String	provenanceOutput

Constructor Summary

Constructors

Constructor and Description
SequenceModel(String name, ModelProvenance provenance, ImmutableFeatureMap featureIDMap, ImmutableOutputInfo<T> outputIDMap)

Method Summary

Modifier and Type	Method and Description
ImmutableFeatureMap	getFeatureIDMap() Gets the feature domain.
String	getName() Gets the model name.
ImmutableOutputInfo<T>	getOutputIDInfo() Gets the output domain.
ModelProvenance	getProvenance()
abstract Map<String,List<com.oracle.labs.mlrg.olcut.util.Pair<String,Double>>>	getTopFeatures(int n) Gets the top n features associated with this model.
List<List<Prediction<T>>>	predict(Iterable<SequenceExample<T>> examples) Uses the model to predict the output for multiple examples.
List<List<Prediction<T>>>	predict(SequenceDataset<T> examples) Uses the model to predict the labels for multiple examples contained in a data set.
abstract List<Prediction<T>>	predict(SequenceExample<T> example) Uses the model to predict the output for a single example.
void	setName(String name) Sets the model name.
static <T extends Output<T>> List<T>	toMaxLabels(List<Prediction<T>> predictions)
String	toString()
boolean	validate(Class<? extends Output<?>> clazz) Validates that this Model does in fact support the supplied output type.

Methods inherited from class java.lang.Object

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

Field Detail

name

protected String name

provenanceOutput

protected final String provenanceOutput

featureIDMap

protected final ImmutableFeatureMap featureIDMap

outputIDMap

protected final ImmutableOutputInfo<T extends Output<T>> outputIDMap

Constructor Detail

SequenceModel

public SequenceModel(String name,
                     ModelProvenance provenance,
                     ImmutableFeatureMap featureIDMap,
                     ImmutableOutputInfo<T> outputIDMap)

Method Detail

validate

public boolean validate(Class<? extends Output<?>> clazz)

Validates that this Model does in fact support the supplied output type.

As the output type is erased at runtime, deserialising a Model is an unchecked operation. This method allows the user to check that the deserialised model is of the appropriate type, rather than seeing if predict(org.tribuo.sequence.SequenceExample<T>) throws a ClassCastException when called.

Parameters:

clazz - The class object to verify the output type against.

Returns:

True if the output type is assignable to the class object type, false otherwise.

getName

public String getName()

Gets the model name.

Returns:

The model name.

setName

public void setName(String name)

Sets the model name.

Parameters:

name - The model name.

getProvenance

public ModelProvenance getProvenance()

Specified by:

getProvenance in interface com.oracle.labs.mlrg.olcut.provenance.Provenancable<ModelProvenance>

toString

public String toString()

Overrides:

toString in class Object

getFeatureIDMap

public ImmutableFeatureMap getFeatureIDMap()

Gets the feature domain.

Returns:

The feature domain.

getOutputIDInfo

public ImmutableOutputInfo<T> getOutputIDInfo()

Gets the output domain.

Returns:

The output domain.

predict

public abstract List<Prediction<T>> predict(SequenceExample<T> example)

Uses the model to predict the output for a single example.

Parameters:

example - the example to predict.

Returns:

the result of the prediction.

predict

public List<List<Prediction<T>>> predict(Iterable<SequenceExample<T>> examples)

Uses the model to predict the output for multiple examples.

Parameters:

examples - the examples to predict.

Returns:

the results of the prediction, in the same order as the examples.

predict

public List<List<Prediction<T>>> predict(SequenceDataset<T> examples)

Uses the model to predict the labels for multiple examples contained in a data set.

Parameters:

examples - the data set containing the examples to predict.

Returns:

the results of the predictions, in the same order as the data set generates the example.

getTopFeatures

public abstract Map<String,List<com.oracle.labs.mlrg.olcut.util.Pair<String,Double>>> getTopFeatures(int n)

Gets the top n features associated with this model.

If the model does not produce per output feature lists, it returns a map with a single element with key Model.ALL_OUTPUTS.

If the model cannot describe it's top features then it returns Collections.emptyMap().

Parameters:

n - the number of features to return. If this value is less than 0, all features should be returned for each class, unless the model cannot score it's features.

Returns:

a map from string outputs to an ordered list of pairs of feature names and weights associated with that feature in the model

toMaxLabels

public static <T extends Output<T>> List<T> toMaxLabels(List<Prediction<T>> predictions)