oracle.dmt.odm.model
Class AttributeImportanceModel

java.lang.Object
  |
  +--oracle.dmt.odm.MiningObject
        |
        +--oracle.dmt.odm.LocatableObject
              |
              +--oracle.dmt.odm.model.MiningModel
                    |
                    +--oracle.dmt.odm.model.AttributeImportanceModel

All Implemented Interfaces:

java.io.Serializable

public abstract class AttributeImportanceModel

extends MiningModel

The abstract class AttributeImportanceModel is a Java representation of the attribute importance model. This class provides methods to access a set of attributes based on their respective importance within the model such as Other statistical metadata.

Attribute Importance is used to select the relevant and important attributes from a build dataset. For ODM 9.2.0, the AttributeImportanceModel is applicable to supervised mining models, and provides an assessment of attribute usefulness. It assigns an importance value and rank to each of the active attributes used for building supervised models. From this ranking, the user can select a subset of important attributes from the complete set of active attributes. The attribute and the corresponding attribute importance value and the rank are stored as a database table.

Since:

9.2.0

See Also:

Serialized Form

Method Summary

AttributeImportanceEntry[]	getAttributeImportanceEntries() Returns all AttributeImportanceEntry objects in the model, each of which contains an attribute name, its importance value, and its rank.
static java.lang.String	getAttributeImportanceTableName(Connection dmsConn, java.lang.String modelName) Returns the table name where each row contains a attribute name, its importance value, and its rank.
java.lang.String[]	getAttributes(AttributeImportanceEntry[] entries) Returns an array of attribute names that correspond to the specified array of AttributeImportanceEntry objects
static java.lang.String[]	getAttributesByImportance(Connection dmsConn, java.lang.String modelName, float threshold, boolean aboveThreshold) Returns an array of attribute names based on the threshold specified for attribute importance value, given the connection to the data mining server and the model name.
static java.lang.String[]	getAttributesByImportance(Connection dmsConn, java.lang.String modelName, float lowerBound, float upperBound) Returns an array of attribute names whose importance values fall in the specified range, given the connection to the data mining server and the model name.
static java.lang.String[]	getAttributesByPercentage(Connection dmsConn, java.lang.String modelName, float percentage, boolean aboveThreshold) Returns an array of attribute names whose importance ranks fall in the specified percentile, given the connection to the data mining server and the model name.
static java.lang.String[]	getAttributesByRank(Connection dmsConn, java.lang.String modelName, int rank, boolean aboveThreshold) Returns an array of attribute names based on the threshold specified for attribute importance rank, given the connection to the data mining server and the model name.
static java.lang.String[]	getAttributesByRank(Connection dmsConn, java.lang.String modelName, int lowerBoundRank, int upperBoundRank) Returns an array of attribute names whose importance ranks fall in the specified range, given the connection to the data mining server and the model name.
static AttributeImportanceEntry[]	getAttributeSetByImportance(Connection dmsConn, java.lang.String modelName, float threshold, boolean aboveThreshold) Returns an array of AttributeImportanceEntry objects based on the threshold specified for attribute importance value, given the connection to the data mining server and the model name.
static AttributeImportanceEntry[]	getAttributeSetByPercentage(Connection dmsConn, java.lang.String modelName, float percentage, boolean aboveThreshold) Returns an array of attribute importance entries whose importance ranks fall in the specified percentile, given the connection to the data mining server and the model name.
static AttributeImportanceEntry[]	getAttributeSetByRank(Connection dmsConn, java.lang.String modelName, int rank, boolean aboveThreshold) Returns an array of attribute importance entries based on the threshold specified for attribute importance rank, given the connection to the data mining server and the model name.
static AttributeImportanceModel	restore(Connection dmsConn, java.lang.String modelName) Returns an instance of AttributeImportanceModel with the specified name persisted in the data mining server given a connection to the data mining server and the model name.

Methods inherited from class oracle.dmt.odm.model.MiningModel

build, getBinCategoricalTableName, getBinNumericTableName, getMiningAlgorithm, getMiningFunction, getMiningFunctionSettings, getModelAlgorithm, getModelMiningFunction, getModelName, getSignature, listContents, listContents, listContents, listContents, remove, rename, storeLocationdata

Methods inherited from class java.lang.Object

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

Method Detail

getAttributeImportanceEntries

public AttributeImportanceEntry[] getAttributeImportanceEntries()

Returns all AttributeImportanceEntry objects in the model, each of which contains an attribute name, its importance value, and its rank.

Returns:

An array of AttributeImportanceEntry objects

getAttributeImportanceTableName

public static java.lang.String getAttributeImportanceTableName(Connection dmsConn,
                                                               java.lang.String modelName)
                                                        throws java.sql.SQLException,
InvalidArgumentException,
MiningObjectException,
ODMException

Returns the table name where each row contains a attribute name, its importance value, and its rank.

Parameters:

dmsConn - The connection to the data mining server

modelName - The name of the model from which the table name is retrieved

Returns:

The name of the model table

Throws:

java.sql.SQLException -

• when the model table does not exist in the database, or
• when the model name is invalid.

InvalidArgumentException -

• when modelName is null, or
• when dmsConn is null.

MiningObjectException - when the model with the specified name does not exist.

ODMException - when modelName is empty or exceeds 64 characters in length.

getAttributeSetByImportance

public static AttributeImportanceEntry[] getAttributeSetByImportance(Connection dmsConn,
                                                                     java.lang.String modelName,
                                                                     float threshold,
                                                                     boolean aboveThreshold)
                                                                                   throws java.sql.SQLException,
InvalidArgumentException,
MiningObjectException,
ODMException

Returns an array of AttributeImportanceEntry objects based on the threshold specified for attribute importance value, given the connection to the data mining server and the model name. The valid value for the threshold must be less than or equal to 1. A negative threshold value implies that the attribute does not give any advantage over the prior information and therefore is irrelevant.

For example, list attributes importance entries with importance values above 0.01.
- threshold = 0.01, aboveThrshold = true

Parameters:

dmsConn - The connection to the data mining server

modelName - The name of the model from which the table name is retrieved

threshold - The value for attribute importance value

aboveThreshold - Returns an array of entries whose importance values are above threashold value if true, or returns an array of entries whose importance values are below threshold if false.

Returns:

An array of AttributeImportanceEntry that meet the specified creteria.

Throws:

java.sql.SQLException -

• when the model table does not exist in the database, or
• when the model name is invalid.

InvalidArgumentException -

• when modelName is null,
• when dmsConn is null, or
• when threshold is greater than 1.

MiningObjectException - when the model with the specified name does not exist.

ODMException - when modelName is empty or exceeds 64 characters in length.

getAttributesByImportance

public static java.lang.String[] getAttributesByImportance(Connection dmsConn,
                                                           java.lang.String modelName,
                                                           float threshold,
                                                           boolean aboveThreshold)
                                                    throws java.sql.SQLException,
InvalidArgumentException,
MiningObjectException,
ODMException

Returns an array of attribute names based on the threshold specified for attribute importance value, given the connection to the data mining server and the model name. The valid value for the threshold must be less than or equal to 1. A negative threshold value implies that the attribute does not give any advantage over the prior information and therefore is irrelevant.

For example, list attributes with importance value above 0.01:
- threshold = 0.01, aboveThrshold = true

Parameters:

dmsConn - The connection to the data mining server

modelName - The name of the model from which the table name is retrieved

threshold - The value for attribute importance value

aboveThreshold - Returns an array of attribute names whose importance values are above threashold value if true, or returns an array of attribute names whose importance values are below threshold if false.

Returns:

An array of attribute names that meet the specified criteria

Throws:

java.sql.SQLException -

• when the model table does not exist in the database,
• when the model name is invalid.

InvalidArgumentException -

• when modelName is null,
• when dmsConn is null, or
• when threshold is greater than 1.

MiningObjectException - when modelName if the model with the specified name does not exist.

ODMException - whenmodelName is empty or exceeds 64 characters in length.

getAttributesByImportance

public static java.lang.String[] getAttributesByImportance(Connection dmsConn,
                                                           java.lang.String modelName,
                                                           float lowerBound,
                                                           float upperBound)
                                                    throws java.sql.SQLException,
InvalidArgumentException,
MiningObjectException,
ODMException

Returns an array of attribute names whose importance values fall in the specified range, given the connection to the data mining server and the model name. The valid value for the threshold must be less than or equal to 1. A negative threshold value implies that the attribute does not give any advantage over the prior information and therefore is irrelevant.

For example, list attributes with importance value between 0.01 and 0.02:
- lowerBound = 0.01, upperBound = 0.02

Parameters:

dmsConn - The connection to the data mining server

modelName - The name of the model from which the table name is retrieved

lowerBound - The lower bound of the threshold

upperBound - The upper bound for the threshold

Returns:

An array of attribute names that meet the specified criteria

Throws:

java.sql.SQLException -

• when the model table does not exist in the database,
• when the model name is invalid.

InvalidArgumentException -

• when modelName is null,
• when dmsConn is null,
• when lowerBound is greater than 1,
• when upperBound is greater than 1, or
• when lowerBound is greater than upperBound.

MiningObjectException - when the model with the specified name does not exist.

ODMException - when modelName is empty or exceeds 64 characters in length.

getAttributeSetByRank

public static AttributeImportanceEntry[] getAttributeSetByRank(Connection dmsConn,
                                                               java.lang.String modelName,
                                                               int rank,
                                                               boolean aboveThreshold)
                                                                             throws java.sql.SQLException,
InvalidArgumentException,
MiningObjectException,
ODMException

Returns an array of attribute importance entries based on the threshold specified for attribute importance rank, given the connection to the data mining server and the model name. The valid value for the threshold must be less than or equal to 1. A negative threshold value implies that the attribute does not give any advantage over the prior information and therefore is irrelevant.

For example, list attribute Importance Entries where rank >= 5:
- threshold = 5, aboveThreshold = true

Parameters:

dmsConn - The connection to the data mining server

modelName - The name of the model from which the table name is retrieved

aboveThreshold - Returns an array of entries whose importance ranks are above rank value if true, or returns an array of entries whose importance ranks are below rank if false.

Returns:

An array of attribute importance entries that meet the specified criteria

Throws:

java.sql.SQLException -

• when the model table does not exist in the database,
• when the model name is invalid.

InvalidArgumentException -

• when modelName is null,
• when dmsConn is null,
• when rank is less than 1.

MiningObjectException - when the model with the specified name does not exist.

ODMException - when modelName is empty or exceeds 64 characters in length.

getAttributesByRank

public static java.lang.String[] getAttributesByRank(Connection dmsConn,
                                                     java.lang.String modelName,
                                                     int rank,
                                                     boolean aboveThreshold)
                                              throws java.sql.SQLException,
InvalidArgumentException,
MiningObjectException,
ODMException

Returns an array of attribute names based on the threshold specified for attribute importance rank, given the connection to the data mining server and the model name. The valid value for the threshold must be less than or equal to 1. A negative threshold value implies that the attribute does not give any advantage over the prior information and therefore is irrelevant.

For example, list attributes where rank >= 5 ordered by rank:
- threshold = 5, aboveThreshold = true

Parameters:

dmsConn - The connection to the data mining server

modelName - The name of the model from which the table name is retrieved

aboveThreshold - Returns an array of entries whose importance ranks are above rank value if true, or returns an array of entries whose importance ranks are below rank if false.

Returns:

An array of attribute names that meet the specified criteria

Throws:

java.sql.SQLException -

• if the model table does not exist in the database, or
• if the model name is invalid.

InvalidArgumentException -

• if modelName is null,
• if dmsConn is null,
• if rank is less than 1.

MiningObjectException - when the model with the specified name does not exist.

ODMException - when modelName is empty or exceeds 64 characters in length.

getAttributesByRank

public static java.lang.String[] getAttributesByRank(Connection dmsConn,
                                                     java.lang.String modelName,
                                                     int lowerBoundRank,
                                                     int upperBoundRank)
                                              throws java.sql.SQLException,
InvalidArgumentException,
MiningObjectException,
ODMException

Returns an array of attribute names whose importance ranks fall in the specified range, given the connection to the data mining server and the model name. The valid value for the ranks ranges between 1 (i.e., the most impoerant) and the number of attributes.

For example, list attributes with rank between 3 and 6:
- lowerBound = 3, upperBound = 6

Parameters:

dmsConn - The connection to the data mining server

modelName - The name of the model from which the table name is retrieved

lowerBoundRank - The lower bound of the ranks

upperBoundRank - The upper bound for the ranks

Returns:

An array of attribute names that meet the specified criteria

Throws:

java.sql.SQLException -

• if the model table does not exist in the database,
• if the model name is invalid.

InvalidArgumentException -

• if modelName is null,
• if dmsConn is null,
• if lowerBoundRank is less than 1,
• if upperBoundRank is less than 1,
• if lowerBoundRank is greater than or equal to upperBoundRank.

MiningObjectException - when the model with the specified name does not exist.

ODMException - when modelName is empty or exceeds 64 characters in length.

getAttributeSetByPercentage

public static AttributeImportanceEntry[] getAttributeSetByPercentage(Connection dmsConn,
                                                                     java.lang.String modelName,
                                                                     float percentage,
                                                                     boolean aboveThreshold)
                                                                                   throws java.sql.SQLException,
InvalidArgumentException,
MiningObjectException,
ODMException

Returns an array of attribute importance entries whose importance ranks fall in the specified percentile, given the connection to the data mining server and the model name. The valid value for the percentage ranges between 0 and 1.

For example, list the attributes importance entries whose rank >= (10% total number of attributes)
- percentage = 0.1, aboveThreshold = false

Parameters:

dmsConn - The connection to the data mining server

modelName - The name of the model from which the table name is retrieved

percentage - The value for percentage of attributes

aboveThreshold - Returns an array of entries whose importance ranks are within the top percentage% if true, or returns an array of entries whose importance ranks are within the bottom percentage% if false.

Returns:

An array of attribute importance entries that meet the specified criteria

Throws:

java.sql.SQLException -

• when the model table does not exist in the database,
• when the model name is invalid.

InvalidArgumentException -

• when modelName is null,
• when dmsConn is null,
• when percentage is less than 0, or ,
• when percentage is greater than 1.

MiningObjectException - when the model with the specified name does not exist.

ODMException - when modelName is empty or exceeds 64 characters in length.

getAttributesByPercentage

public static java.lang.String[] getAttributesByPercentage(Connection dmsConn,
                                                           java.lang.String modelName,
                                                           float percentage,
                                                           boolean aboveThreshold)
                                                    throws java.sql.SQLException,
InvalidArgumentException,
MiningObjectException,
ODMException

Returns an array of attribute names whose importance ranks fall in the specified percentile, given the connection to the data mining server and the model name. The valid value for the percentage ranges between 0 and 1.

For example, list the attributes names whose rank >= (10% total number of attributes)
- percentage = 0.1, aboveThreshold = false

Parameters:

dmsConn - The connection to the data mining server

modelName - The name of the model from which the table name is retrieved

percentage - The value for percentage of attributes

aboveThreshold - Returns an array of attribute names whose importance ranks are within the top percentage% if true, or returns an array of attribute names whose importance ranks are within the bottom percentage% if false.

Returns:

An array of attribute importance entries that meet the specified criteria

Throws:

java.sql.SQLException -

• when the model table does not exist in the database,
• when the model name is invalid.

InvalidArgumentException -

• when modelName is null,
• when dmsConn is null,
• when percentage is less than 0, or ,
• when percentage is greater than 1.

MiningObjectException - when the model with the specified name does not exist.

ODMException - when modelName is empty or exceeds 64 characters in length.

getAttributes

public java.lang.String[] getAttributes(AttributeImportanceEntry[] entries)

Returns an array of attribute names that correspond to the specified array of AttributeImportanceEntry objects

Parameters:

entries - An array of AttributeImportanceEntry objects

Returns:

An array of attribute names

restore

public static AttributeImportanceModel restore(Connection dmsConn,
                                               java.lang.String modelName)
                                                             throws java.sql.SQLException,
MiningObjectException,
InvalidArgumentException,
ODMException

Returns an instance of AttributeImportanceModel with the specified name persisted in the data mining server given a connection to the data mining server and the model name.

Parameters:

dmsConn - The connection to the data mining server

modelName - The name of the model to be restored

Returns:

An instance of AttributeImportanceModel restored

Throws:

InvalidArgumentException -

• when modelName is null,
• when dmsConn is null

ODMException - when modelName is empty or exceeds 64 acharacter in length.

java.sql.SQLException -

• when the model table does not exist in the database,
• when the model name is invalid

MiningObjectException - when the restored model is not an instance of AttributeImportanceModel.