oracle.dmt.odm.settings.function
Class ClassificationFunctionSettings

java.lang.Object
  |
  +--oracle.dmt.odm.MiningObject
        |
        +--oracle.dmt.odm.LocatableObject
              |
              +--oracle.dmt.odm.settings.function.MiningFunctionSettings
                    |
                    +--oracle.dmt.odm.settings.function.SupervisedFunctionSettings
                          |
                          +--oracle.dmt.odm.settings.function.ClassificationFunctionSettings

All Implemented Interfaces:

java.io.Serializable

public class ClassificationFunctionSettings

extends SupervisedFunctionSettings

An instance of ClassificationFunctionSettings describes the settings necessary to build a classification model.

Given a pre-determined set of class values in a target attribute, classification analyzes a training dataset to determine to which class a given data record belongs. Classification has been used in customer segmentation, business modeling, and credit analysis.

In ODM 9.2.0, two classification algorithms are supported: Naive Bayes and Adaptive Bayes Network (decision tree). If no algorithm is specified, the Naive Bayes algorithm is chosen by default.

See Also:

AdaptiveBayesNetworkSettings, NaiveBayesSettings, Serialized Form

Constructor Summary

ClassificationFunctionSettings(float accuracy, LogicalDataSpecification lds, DataUsageSpecification dus)
Deprecated. As of ODM 9.2.0. The parameter accuracy is not supported. Use other constructor without accuracy parameter.

ClassificationFunctionSettings(float accuracy, LogicalDataSpecification lds, DataUsageSpecification dus, MiningAlgorithmSettings algorithm)
Deprecated. As of ODM 9.2.0. The parameter accuracy is not supported. Use other constructor without accuracy parameter.

ClassificationFunctionSettings(LogicalDataSpecification lds, DataUsageSpecification dus)
Creates an instance of ClassificationFunctionSettings object.

ClassificationFunctionSettings(LogicalDataSpecification lds, DataUsageSpecification dus, MiningAlgorithmSettings algorithm)
Creates an instance of ClassificationFunctionSettings object.

Method Summary

static ClassificationFunctionSettings	create(Connection dmsConn, MiningAlgorithmSettings algorithm, PhysicalDataSpecification pdsDataSpec, java.lang.String sTargetAttrName, AttributeType atTargetAttributeType, DataPreparationStatus dpStatus) Creates an instance of ClassificationFunctionSettings from the specified parameters.
CategoryMatrix	getCostMatrix() Returns the cost matrix set in this mining function settings.
PriorProbabilities	getPriors() Returns the prior probabilities of the input data set in this function settings.
MiningAttribute	getTargetAttribute() Returns the target attribute set in this functional settings.
void	setCostMatrix(CategoryMatrix costMatrix) Sets the cost matrix to be used for test operation.
void	setPriors(PriorProbabilities prior) Sets the prior probabilities of the input data.
void	validate() Validates the contents of this mining function settings.

Methods inherited from class oracle.dmt.odm.settings.function.SupervisedFunctionSettings

getDesiredAccuracy, getPredictorAttributes, getTargetAttributes, getUnpreparedPredictorAttributes, setDesiredAccuracy

Methods inherited from class oracle.dmt.odm.settings.function.MiningFunctionSettings

activateAll, adjustAttributesType, adjustAttributeUsage, adjustAttributeUsage, adjustDataPreparationStatus, adjustSparsitySpecification, getDataUsageSpecification, getLogicalDataSpecification, getMiningAlgorithmName, getMiningAlgorithmSettings, getMiningAttributes, getMiningFunction, getName, getSettingsID, getSupplementalAttributes, getValidateBeforeStore, listContents, listContents, listContents, listContents, remove, rename, restore, restore, restorePastVersion, restorePastVersion, setDataUsageSpecification, setLogicalDataSpecification, setMiningAlgorithmName, setMiningAlgorithmSettings, setMiningFunction, setValidateBeforeStore, store

Methods inherited from class java.lang.Object

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

Constructor Detail

ClassificationFunctionSettings

public ClassificationFunctionSettings(float accuracy,
LogicalDataSpecification lds,
DataUsageSpecification dus,
MiningAlgorithmSettings algorithm)
                               throws InvalidArgumentException

Deprecated. As of ODM 9.2.0. The parameter accuracy is not supported. Use other constructor without accuracy parameter.

Creates an instance of ClassificationFunctionSettings object.

Parameters:

accuracy - The desired minimum accuracy of the classification model to be built

lds - An instance of LogicalDataSpecification that specifies the logical representation of the input data used for model building

dus - An instance of DataUsageSpecification that defines how the logical data is to be used for model building

algorithm - An instance of MiningAlgorithmSettings that specifies a classification algorithm to be used for model building. If null, a default algorithm is chosen by the DMS.

Throws:

InvalidArgumentException -

• if lds is null, or
• if accuracy is invalid (less than 0 or greater than 1).

ClassificationFunctionSettings

public ClassificationFunctionSettings(LogicalDataSpecification lds,
DataUsageSpecification dus,
MiningAlgorithmSettings algorithm)
                               throws InvalidArgumentException

Creates an instance of ClassificationFunctionSettings object.

Parameters:

lds - An instance of LogicalDataSpecification that specifies the logical representation of the input data used for model building

dus - An instance of DataUsageSpecification that defines how the logical data is to be used for model building

algorithm - An instance of MiningAlgorithmSettings that specifies a classification algorithm to be used for model building. If null, a default algorithm is chosen by the DMS.

Throws:

InvalidArgumentException - when lds is null.

Since:

9.2.0

ClassificationFunctionSettings

public ClassificationFunctionSettings(float accuracy,
LogicalDataSpecification lds,
DataUsageSpecification dus)
                               throws InvalidArgumentException

Deprecated. As of ODM 9.2.0. The parameter accuracy is not supported. Use other constructor without accuracy parameter.

Creates an instance of ClassificationFunctionSettings object. A default algorithm settings is selected by the DMS, and applied to model building.

Parameters:

accuracy - The desired minimum accuracy of the classification model to be built

lds - An instance of LogicalDataSpecification that specifies the logical representation of the input data used for model building

dus - An instance of DataUsageSpecification that defines how the logical data is to be used for model building

Throws:

InvalidArgumentException -

• if lds is null, or
• if accuracy is invalid (less than 0 or greater than 1).

ClassificationFunctionSettings

public ClassificationFunctionSettings(LogicalDataSpecification lds,
DataUsageSpecification dus)
                               throws InvalidArgumentException

Creates an instance of ClassificationFunctionSettings object. A default algorithm settings is selected by the DMS, and applied to model building.

Parameters:

lds - An instance of LogicalDataSpecification that specifies the logical representation of the input data used for model building

dus - An instance of DataUsageSpecification that defines how the logical data is to be used for model building

Throws:

InvalidArgumentException -

• if lds is null, or
• if accuracy is invalid (less than 0 or greater than 1).

Since:

9.2.0

Method Detail

getTargetAttribute

public MiningAttribute getTargetAttribute()
                                                       throws MiningObjectException

Returns the target attribute set in this functional settings.

Returns:

The target mining attribute.

MiningObjectException

See Also:

AttributeUsage

validate

public void validate()
              throws MiningObjectException

Validates the contents of this mining function settings.

Throws:

MiningObjectException - when this mining function settings is invalid

setPriors

public void setPriors(PriorProbabilities prior)
               throws InvalidArgumentException

Sets the prior probabilities of the input data.

Parameters:

prior - The prior probabilities to be set

Throws:

InvalidArgumentException - when prior is null.

Since:

9.2.0

See Also:

PriorProbabilities

getPriors

public PriorProbabilities getPriors()

Returns the prior probabilities of the input data set in this function settings.

Returns:

The prior probabilities

Since:

9.2.0

See Also:

PriorProbabilities

setCostMatrix

public void setCostMatrix(CategoryMatrix costMatrix)
                   throws InvalidArgumentException

Sets the cost matrix to be used for test operation.

Parameters:

costMatrix - - The cost matrix to be set

Throws:

InvalidArgumentException - when costMatrix is null.

Since:

9.2.0

See Also:

CategoryMatrix

getCostMatrix

public CategoryMatrix getCostMatrix()

Returns the cost matrix set in this mining function settings.

Returns:

The cost matrix

Since:

9.2.0

See Also:

CategoryMatrix

create

public static ClassificationFunctionSettings create(Connection dmsConn,
MiningAlgorithmSettings algorithm,
PhysicalDataSpecification pdsDataSpec,
                                                    java.lang.String sTargetAttrName,
AttributeType atTargetAttributeType,
DataPreparationStatus dpStatus)
                                                                              throws java.sql.SQLException,
InvalidArgumentException,
MiningObjectException,
ODMException

Creates an instance of ClassificationFunctionSettings from the specified parameters.

Parameters:

dmsConn - A connection to the data mining server

algorithm - An instance of MiningAlgorithmSettings

pdsDataSpec - An instance of PhysicalDataSpecification that specifies the input data

sTargetAttrName - The name of the target attribute

atTargetAttributeType - The type of the target attribute. Only string or integer type is allowed.

dpStatus - An instance of DataPreparationStatus that indicates whether the data is discretized.

Returns:

ClassificationFunctionSettings - A created instance of classification function settings

Throws:

InvalidArgumentException -

• if dmsConn is null,
• if pdsDataSpec is null,
• if sTargetAttrName is null,
• if atTargetAttributeType is null, or
• if atTargetAttributeType is neither string nor integer.

ODMException - when sTargetAttrName is empty or exceeds 30 bytes in length

java.sql.SQLException - when an error occurs during access to the input data

MiningObjectException - inherited exception