[Package Index] [Index] [Users' Guide] [SVAPI Home Page]

Interface COM.novell.SVAPI.ModelBasedEngine

public interface ModelBasedEngine extends Engine

The ModelBasedEngine interface is the interface off of which all engine which use model's inherit.



Method Index

createModel()
Creates a temporary model.
createModel(Model[])
Creates a temporary model based on an array of models.
createModel(String)
Creates a model.
createModel(String, Model[])
Creates a model based on an array of models.
destroyModel(String)
Destroys a model from permanent storage.
getAllModels()
Returns an enumeration of the names of all persistent models.
getEqualErrorPoint()
Returns the equal error point.
getModel(String)
Gets a model by name.
getThreshold()
Returns an engine's threshold value.
isModelNameValid(String)
Returns true if the name is a valid name for a model.
setConfidenceThreshold(int)
Sets an engine's threshold from a confidence level.
setRawThreshold(double)
Sets an engine's threshold.
verifyText(Model, Utterance, String)
Verifies an utterance against a string to determine whether the speaker spoke the prompted text.



Methods



getModel

public abstract Model getModel(String ModelName)
Gets a model by name. This method is used to reinstantiate a model that was previously created. If the model is already instantiated, this function should return the same object.
Exceptions:
See Also: dispose


createModel

public abstract Model createModel()
Creates a temporary model. The model will need to be trained before it can be used.
Exceptions:
See Also: dispose, copy


createModel

public abstract Model createModel(String Name)
Creates a model. The model will need to be trained before it can be used.
Exceptions:
See Also: isModelNameValid, dispose, copy


createModel

public abstract Model createModel(Model From[])
Creates a temporary model based on an array of models.
Exceptions:
See Also: dispose, copy


createModel

public abstract Model createModel(String Name, Model From[])
Creates a model based on an array of models.
Exceptions:
See Also: isModelNameValid, dispose, copy


destroyModel

public abstract void destroyModel(String Name)
Destroys a model from permanent storage. A cohort-based system may keep the voice information and simply make the model unreferencable by name.
Exceptions:
See Also: dispose


getAllModels

public abstract RemoteList getAllModels()
Returns an enumeration of the names of all persistent models.
Exceptions:
See Also: createModel


isModelNameValid

public abstract boolean isModelNameValid(String Name)
Returns true if the name is a valid name for a model. A valid model name consists of a string of alphanumerics and spaces that is between 1 and 127 characters in length. An engine by necessity may need to reserve a portion of the name space for internal data storage and would reject such names. It is recommended that if an engine need to reserve names, that it prefix all names with some non-obvious string, such as a space followed by the engine name. Model names are case sensitive.
Exceptions:


verifyText

public abstract int verifyText(Model Model, Utterance Utterance, String Text)
Verifies an utterance against a string to determine whether the speaker spoke the prompted text. The value returned is a confidence level and must be between MIN_CONFIDENCE and MAX_CONFIDENCE.
Exceptions:


setRawThreshold

public abstract void setRawThreshold(double RawScore)
Sets an engine's threshold. When an engine gets a raw score, it is compared against the threshold value to get a boolean acceptance score.
Exceptions:
See Also: setConfidenceThreshold, Score


setConfidenceThreshold

public abstract void setConfidenceThreshold(int ConfidenceLevel)
Sets an engine's threshold from a confidence level. The engine maps the confidence level to a raw score and sets its threshold to that value. When an engine gets a raw score, it is compared against the threshold value to get a boolean acceptance score. The confidence level must be between MIN_CONFIDENCE and MAX_CONFIDENCE.
Exceptions:
See Also: setRawThreshold, Score


getThreshold

public abstract double getThreshold()
Returns an engine's threshold value.
Exceptions:
See Also: setRawThreshold, setConfidenceThreshold, Score


getEqualErrorPoint

public abstract double getEqualErrorPoint()
Returns the equal error point.
Exceptions:
See Also: setRawThreshold


Variables



MAX_CONFIDENCE

public final static int MAX_CONFIDENCE
The maximum confidence level. Confidence levels are used in text verification, threshold setting and in the ConfidenceMapping class.
See Also: verifyText, setConfidenceThreshold, ConfidenceMapping


MIN_CONFIDENCE

public final static int MIN_CONFIDENCE
The minimum confidence level. Confidence levels are used in text verification, threshold setting and in the ConfidenceMapping class.
See Also: verifyText, setConfidenceThreshold, ConfidenceMapping



[Package Index] [Index] [Users' Guide] [SVAPI Home Page]

rwille@novell.com