org.knime.base.node.mine.sota.logic
Class SotaManager

java.lang.Object
  org.knime.base.node.mine.sota.logic.SotaManager

public class SotaManager
extends Object

Author:

Kilian Thiel, University of Konstanz

Field Summary

static double	LR_ANCESTOR Default value of the ancestor learningrate.
static double	LR_ANCESTOR_MAX Maximal value of the ancestor learningrate.
static double	LR_ANCESTOR_MIN Minimal value of the ancestor learningrate.
static double	LR_SISTER Default value of the sister learningrate.
static double	LR_SISTER_MAX Maximal value of the ancestor learningrate.
static double	LR_SISTER_MIN Minimal value of the ancestor learningrate.
static double	LR_WINNER Default value of the Winner learningrate.
static double	LR_WINNER_MAX Maximal value of the Winner learningrate.
static double	LR_WINNER_MIN Minimal value of the Winner learningrate.
static double	MIN_ERROR Default value of minimal error.
static double	MIN_ERROR_MAX Maximal value of minimal error.
static double	MIN_ERROR_MIN Minimal value of minimal error.
static double	MIN_RESOURCE Default value of minimal resource.
static double	MIN_RESOURCE_MAX Maximal value of minimal resource.
static double	MIN_RESOURCE_MIN Minimal value of minimal resource.
static double	MIN_VARIABILITY Default value of minimal variability.
static double	MIN_VARIABILITY_MAX Maximal value of minimal variability.
static double	MIN_VARIABILITY_MIN Minimal value of minimal variability.
static boolean	USE_HIERARCHICAL_FUZZY_DATA Is hierarchical fuzzy data used or not .
static boolean	USE_VARIABILITY Default flag setting.

Constructor Summary

SotaManager()
Creates new instance of SotaManager with default settings.

Method Summary

double	doCycle() Computes one cycle of the sota algorithm, does the spilt afterwards and returns the variability.
double	doEpoch() Computes one epoch of the sota algorithm and retunrs the error of the tree after this epoch.
double	doTraining() Trains the tree as many cycles as it takes to reduce the variability value to the given minimum and returns the variability value.
static void	getCells(ArrayList<SotaTreeCell> cells, SotaTreeCell currentCell) Collects all cells of the tree recursive.
double	getCycle()
int	getDimension()
String	getDistance()
double	getEpoch()
ExecutionMonitor	getExecutionMonitor()
DataArray	getInDataContainer()
double	getLearningrateAncestor()
double	getLearningrateSister()
double	getLearningrateWinner()
int	getMaxHierarchicalLevel()
double	getMinError()
double	getMinResource()
double	getMinVariability()
DataArray	getOriginalData() Returns the original DataTableSpec.
SotaTreeCell	getRoot()
static void	initDefaultLocations(HashMap<SotaTreeCell,SotaTreeCellLocations> cellLocations, SotaTreeCell currentCell)
void	initializeTree(DataTable inData, DataArray originalData, ExecutionMonitor exec, int indexOfClassColumn) Initializes the tree by creating the root node and two children cells of the root node.
boolean	isTrained()
boolean	isUseHierarchicalFuzzyData()
boolean	isUseVariability()
void	readSettings(NodeSettingsRO settings, boolean validateOnly) Reads settings out of given NodeSettings object and validates it.
void	reset() Resets the SotaManager.
void	saveSettingsTo(NodeSettingsWO settings) Saves settings of algorithm to given NodeSettings object.
void	setDistance(String distance)
void	setInData(DataArray inData) Sets in data.
void	setLearningrateAncestor(double ancestor)
void	setLearningrateSister(double sister)
void	setLearningrateWinner(double winner)
void	setMaxHierarchicalLevel(int maxLevel) Sets the maximum heirarchical level.
void	setMinError(double error)
void	setMinResource(double resource)
void	setMinVariability(double variability)
void	setOriginalData(DataArray origData) Sets original data.
void	setRoot(SotaTreeCell root) Sets the root node.
void	setUseHierarchicalFuzzyData(boolean hierarchicalFuzzyData)
void	setUseVariability(boolean variability)

Methods inherited from class java.lang.Object

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

Field Detail

LR_WINNER

public static final double LR_WINNER

Default value of the Winner learningrate.

See Also:

Constant Field Values

LR_WINNER_MIN

public static final double LR_WINNER_MIN

Minimal value of the Winner learningrate.

See Also:

Constant Field Values

LR_WINNER_MAX

public static final double LR_WINNER_MAX

Maximal value of the Winner learningrate.

See Also:

Constant Field Values

LR_ANCESTOR

public static final double LR_ANCESTOR

Default value of the ancestor learningrate.

See Also:

Constant Field Values

LR_ANCESTOR_MIN

public static final double LR_ANCESTOR_MIN

Minimal value of the ancestor learningrate.

See Also:

Constant Field Values

LR_ANCESTOR_MAX

public static final double LR_ANCESTOR_MAX

Maximal value of the ancestor learningrate.

See Also:

Constant Field Values

LR_SISTER

public static final double LR_SISTER

Default value of the sister learningrate.

See Also:

Constant Field Values

LR_SISTER_MIN

public static final double LR_SISTER_MIN

Minimal value of the ancestor learningrate.

See Also:

Constant Field Values

LR_SISTER_MAX

public static final double LR_SISTER_MAX

Maximal value of the ancestor learningrate.

See Also:

Constant Field Values

MIN_VARIABILITY

public static final double MIN_VARIABILITY

Default value of minimal variability.

See Also:

Constant Field Values

MIN_VARIABILITY_MIN

public static final double MIN_VARIABILITY_MIN

Minimal value of minimal variability.

See Also:

Constant Field Values

MIN_VARIABILITY_MAX

public static final double MIN_VARIABILITY_MAX

Maximal value of minimal variability.

See Also:

Constant Field Values

MIN_RESOURCE

public static final double MIN_RESOURCE

Default value of minimal resource.

See Also:

Constant Field Values

MIN_RESOURCE_MIN

public static final double MIN_RESOURCE_MIN

Minimal value of minimal resource.

See Also:

Constant Field Values

MIN_RESOURCE_MAX

public static final double MIN_RESOURCE_MAX

Maximal value of minimal resource.

See Also:

Constant Field Values

MIN_ERROR

public static final double MIN_ERROR

Default value of minimal error.

See Also:

Constant Field Values

MIN_ERROR_MIN

public static final double MIN_ERROR_MIN

Minimal value of minimal error.

See Also:

Constant Field Values

MIN_ERROR_MAX

public static final double MIN_ERROR_MAX

Maximal value of minimal error.

See Also:

Constant Field Values

USE_VARIABILITY

public static final boolean USE_VARIABILITY

Default flag setting.

See Also:

Constant Field Values

USE_HIERARCHICAL_FUZZY_DATA

public static final boolean USE_HIERARCHICAL_FUZZY_DATA

Is hierarchical fuzzy data used or not .

See Also:

Constant Field Values

Constructor Detail

SotaManager

public SotaManager()

Creates new instance of SotaManager with default settings.

Method Detail

reset

public void reset()

Resets the SotaManager.

initializeTree

public void initializeTree(DataTable inData,
                           DataArray originalData,
                           ExecutionMonitor exec,
                           int indexOfClassColumn)
                    throws CanceledExecutionException

Initializes the tree by creating the root node and two children cells of the root node. The nodes data are the mean values of the input data rows.

Parameters:

inData - the table with the input data

originalData - the original data

exec - the execution monitor to set

indexOfClassColumn - The index of the column containing the class information. If value is -1 class values are ignored.

Throws:

CanceledExecutionException - if user canceled the process

doTraining

public double doTraining()
                  throws CanceledExecutionException

Trains the tree as many cycles as it takes to reduce the variability value to the given minimum and returns the variability value.

Returns:

the variability value

Throws:

CanceledExecutionException - if training has been canceled

doCycle

public double doCycle()
               throws CanceledExecutionException

Computes one cycle of the sota algorithm, does the spilt afterwards and returns the variability.

Returns:

the variability of the tree after this cycle

Throws:

CanceledExecutionException - if user canceled the process

doEpoch

public double doEpoch()
               throws CanceledExecutionException

Computes one epoch of the sota algorithm and retunrs the error of the tree after this epoch.

Returns:

the error of the tree after this epoch

Throws:

CanceledExecutionException - is execution was canceled.

getCells

public static void getCells(ArrayList<SotaTreeCell> cells,
                            SotaTreeCell currentCell)

Collects all cells of the tree recursive.

Parameters:

cells - the ArrayList to store the cells in

currentCell - the current cell to check

initDefaultLocations

public static void initDefaultLocations(HashMap<SotaTreeCell,SotaTreeCellLocations> cellLocations,
                                        SotaTreeCell currentCell)

getLearningrateAncestor

public double getLearningrateAncestor()

Returns:

the learningrateAncestor

setLearningrateAncestor

public void setLearningrateAncestor(double ancestor)

Parameters:

ancestor - the learningrateAncestor to set

getLearningrateSister

public double getLearningrateSister()

Returns:

the learningrateSister

setLearningrateSister

public void setLearningrateSister(double sister)

Parameters:

sister - the learningrateSister to set

getLearningrateWinner

public double getLearningrateWinner()

Returns:

the learningrateWinner

setLearningrateWinner

public void setLearningrateWinner(double winner)

Parameters:

winner - the learningrateWinner to set

getMinError

public double getMinError()

Returns:

the minError

setMinError

public void setMinError(double error)

Parameters:

error - the minError to set

getMinResource

public double getMinResource()

Returns:

the minResource

setMinResource

public void setMinResource(double resource)

Parameters:

resource - the minResource to set

getMinVariability

public double getMinVariability()

Returns:

the minVariability

setMinVariability

public void setMinVariability(double variability)

Parameters:

variability - the minVariability to set

getDimension

public int getDimension()

Returns:

the dimension

getInDataContainer

public DataArray getInDataContainer()

Returns:

the inDataContainer

getRoot

public SotaTreeCell getRoot()

Returns:

the root

getCycle

public double getCycle()

Returns:

the cycle

getEpoch

public double getEpoch()

Returns:

the epoch

getExecutionMonitor

public ExecutionMonitor getExecutionMonitor()

Returns:

the execution monitor

isUseVariability

public boolean isUseVariability()

Returns:

the useVariability

setUseVariability

public void setUseVariability(boolean variability)

Parameters:

variability - the useVariability to set

getDistance

public String getDistance()

Returns:

the distance

setDistance

public void setDistance(String distance)

Parameters:

distance - the distance to set

isUseHierarchicalFuzzyData

public boolean isUseHierarchicalFuzzyData()

Returns:

the useHierarchicalFuzzyData

setUseHierarchicalFuzzyData

public void setUseHierarchicalFuzzyData(boolean hierarchicalFuzzyData)

Parameters:

hierarchicalFuzzyData - the useHierarchicalFuzzyData to set

isTrained

public boolean isTrained()

Returns:

the trained

getMaxHierarchicalLevel

public int getMaxHierarchicalLevel()

Returns:

the maxHierarchicalLevel

saveSettingsTo

public void saveSettingsTo(NodeSettingsWO settings)

Saves settings of algorithm to given NodeSettings object.

Parameters:

settings - NodeSettings object to store settings in

readSettings

public void readSettings(NodeSettingsRO settings,
                         boolean validateOnly)
                  throws InvalidSettingsException

Reads settings out of given NodeSettings object and validates it. If validateOnly is false, NodeSettings data will be stored in algorithm too.

Parameters:

settings - NodeSettings object to get settings from

validateOnly - if true, settings will be validated only, if false, settings will be stored in algorithm

Throws:

InvalidSettingsException - will be thrown if given settings are not valid

getOriginalData

public DataArray getOriginalData()

Returns the original DataTableSpec.

Returns:

the original DataTableSpec

setMaxHierarchicalLevel

public void setMaxHierarchicalLevel(int maxLevel)

Sets the maximum heirarchical level.

Parameters:

maxLevel - the maximum heirarchical level to set

setRoot

public void setRoot(SotaTreeCell root)

Sets the root node.

Parameters:

root - the root node to set

setInData

public void setInData(DataArray inData)

Sets in data.

Parameters:

inData - in data to set

setOriginalData

public void setOriginalData(DataArray origData)

Sets original data.

Parameters:

origData - original data to set