org.knime.base.node.mine.decisiontree2.model
Class DecisionTreeNode

java.lang.Object
  org.knime.base.node.mine.decisiontree2.model.DecisionTreeNode

All Implemented Interfaces:

Serializable, TreeNode

Direct Known Subclasses:

DecisionTreeNodeLeaf, DecisionTreeNodeSplit

public abstract class DecisionTreeNode
extends Object
implements TreeNode, Serializable

The base abstract implementations of a node of a decision tree. Separate implementations for a leaf and a split node (abstract) exist.

Author:

Michael Berthold, University of Konstanz, Christoph Sieb, University of Konstanz

See Also:

Serialized Form

Constructor Summary

(package private)	DecisionTreeNode() Empty Constructor visible only within package.
protected	DecisionTreeNode(int nodeId, DataCell majorityClass, LinkedHashMap<DataCell,Double> classCounts) Constructor of base class.
protected	DecisionTreeNode(Node xmlNode, DataCellStringMapper mapper) Constructor of base class.

Method Summary

protected void	addColorToMap(Color col, double weight) Adds the given color to the color map.
abstract void	addCoveredColor(DataRow row, DataTableSpec spec, double weight) Add colors for a row of values if they fall within a specific node/branch.
abstract void	addCoveredPattern(DataRow row, DataTableSpec spec, double weight) Add patterns given as a row of values if they fall within a specific node.
abstract boolean	addNodeToTreeDepthFirst(DecisionTreeNode node, int ix) Add a new node to the tree structure based on a depth-first indexing strategy.
abstract Enumeration<DecisionTreeNode>	children()
HashMap<Color,Double>	coveredColors()
abstract Set<RowKey>	coveredPattern()
static DecisionTreeNode	createNewNode(Node xmlNode, DataCellStringMapper mapper) Create new node from XML-information.
static DecisionTreeNode	createNodeFromPredictorParams(ModelContentRO predParams, DecisionTreeNode parent) Creates a new DecisionTreeNode (and all it's children!) based on an model content object.
abstract boolean	getAllowsChildren()
abstract DecisionTreeNode	getChildAt(int pos)
abstract int	getChildCount()
LinkedHashMap<DataCell,Double>	getClassCounts() Return class counts, that is how many patterns (also fractions of) for each class were encountered in this branch during training.
abstract LinkedHashMap<DataCell,Double>	getClassCounts(DataRow row, DataTableSpec spec) Determine class counts for a new pattern given as a row of values.
abstract int	getCountOfSubtree() Returns the count of the subtree.
Object	getCustomData() To get the custom data object.
double	getEntireClassCount() Return number of patterns of all classes.
abstract int	getIndex(TreeNode node) Returns the index of node in the receivers children.
DataCell	getMajorityClass() Return majority class of this node.
(package private) double	getOverallColorCount()
double	getOwnClassCount() Return number of patterns of correct class (= majority class in a non-risk decision tree).
int	getOwnIndex()
DecisionTreeNode	getParent()
String	getPrefix() Returns the prefix of this node representing the condition.
abstract String	getStringSummary()
Pair<DataCell,LinkedHashMap<DataCell,Double>>	getWinnerAndClasscounts(DataRow row, DataTableSpec spec) Classify a new pattern given as a row of values.
abstract boolean	isLeaf()
void	loadFromPredictorParams(ModelContentRO predParams) Load node from a model content object.
abstract void	loadNodeInternalsFromPredParams(ModelContentRO pConf) Load internal node settings from a model content object.
(package private) boolean	newColors()
void	resetColorInformation() Clean all color information in this node and all children.
abstract void	saveNodeInternalsToPredParams(ModelContentWO pConf, boolean saveKeysAndPatterns) Save internal node settings to a model content object.
void	saveToPredictorParams(ModelContentWO predParams, boolean saveColorsAndKeys) Save node to a model content object.
void	setCoveredColors(HashMap<Color,Double> coveredColors) Sets the covered colors distribution directly.
void	setCustomData(Object customData) To set a custom data object.
void	setParent(DecisionTreeNode parent) Set parent of this node.
void	setPrefix(String pf) Set information about this node, e.g.
String	toString()

Methods inherited from class java.lang.Object

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

Constructor Detail

DecisionTreeNode

DecisionTreeNode()

Empty Constructor visible only within package.

DecisionTreeNode

protected DecisionTreeNode(Node xmlNode,
                           DataCellStringMapper mapper)

Constructor of base class. Read all type-invariant information from XML file.

Parameters:

xmlNode - XML node object

mapper - map translating column names to DataCells and vice versa

DecisionTreeNode

protected DecisionTreeNode(int nodeId,
                           DataCell majorityClass,
                           LinkedHashMap<DataCell,Double> classCounts)

Constructor of base class. The necessary data is provided directly in the constructor.

Parameters:

nodeId - the id of this node

majorityClass - the majority class of the records in this node

classCounts - the class distribution of the data in this node

Method Detail

createNewNode

public static DecisionTreeNode createNewNode(Node xmlNode,
                                             DataCellStringMapper mapper)

Create new node from XML-information. Note that this constructor only constructs the node itself and does not generate any other nodes connected to it - it will solely read it's children's indices from the XML file. This function serves as a factory for subclasses that are distinguished here - the constructors of these classes are called and will read their individual information themselves. Most "node" type stuff is handled here, however.

Parameters:

xmlNode - XML Information for this node

mapper - map translating column names to DataCells and vice versa

Returns:

new node initialized from XML node or null if type is not recognized

addNodeToTreeDepthFirst

public abstract boolean addNodeToTreeDepthFirst(DecisionTreeNode node,
                                                int ix)

Add a new node to the tree structure based on a depth-first indexing strategy.

Parameters:

node - node to be inserted

ix - index of this node in depth first traversal order

Returns:

true only if the node was successfully inserted

getMajorityClass

public DataCell getMajorityClass()

Return majority class of this node.

Returns:

majority class

getClassCounts

public LinkedHashMap<DataCell,Double> getClassCounts()

Return class counts, that is how many patterns (also fractions of) for each class were encountered in this branch during training.

Returns:

class counts

getWinnerAndClasscounts

public final Pair<DataCell,LinkedHashMap<DataCell,Double>> getWinnerAndClasscounts(DataRow row,
                                                                                   DataTableSpec spec)
                                                                            throws Exception

Classify a new pattern given as a row of values. Returns the winning class and the class counts of all classes.

Parameters:

row - input pattern

spec - the corresponding table spec

Returns:

class of pattern the decision tree predicts

Throws:

Exception - if something went wrong (unknown attribute for example)

getClassCounts

public abstract LinkedHashMap<DataCell,Double> getClassCounts(DataRow row,
                                                              DataTableSpec spec)

Determine class counts for a new pattern given as a row of values. Returns a HashMap listing counts for all classes.

Parameters:

row - input pattern

spec - the corresponding table spec

Returns:

HashMap class/count

getOwnClassCount

public double getOwnClassCount()

Return number of patterns of correct class (= majority class in a non-risk decision tree). Note that there can be fractions of patterns, for example when we have built a fuzzy decision tree or there were missing values.

Returns:

number (and fractions) of patterns of class of this node

getEntireClassCount

public double getEntireClassCount()

Return number of patterns of all classes. Note that there can be fractions of patterns, for example when we have built a fuzzy decision tree or there were missing values.

Returns:

number (and fractions) of patterns of all classes

addCoveredPattern

public abstract void addCoveredPattern(DataRow row,
                                       DataTableSpec spec,
                                       double weight)
                                throws Exception

Add patterns given as a row of values if they fall within a specific node. Usually only Leafs will actually hold a list of RowKeys, all intermediate nodes will collect "their" information recursively.

Parameters:

row - input pattern

spec - the corresponding table spec

weight - the weight of the row (between 0.0 and 1.0)

Throws:

Exception - if something went wrong (unknown attribute for example)

addCoveredColor

public abstract void addCoveredColor(DataRow row,
                                     DataTableSpec spec,
                                     double weight)
                              throws Exception

Add colors for a row of values if they fall within a specific node/branch. Used if we don't want to (or can't anymore) store the pattern itself. We still want the color pie chart to be correct.

Parameters:

row - input pattern

spec - the corresponding table spec

weight - the weight of the row (between 0.0 and 1.0)

Throws:

Exception - if something went wrong (unknown attribute for example)

coveredPattern

public abstract Set<RowKey> coveredPattern()

Returns:

set of data cells which are the row keys that are covered by all nodes of this branch

coveredColors

public final HashMap<Color,Double> coveredColors()

Returns:

list of colors and coverage counts covered by this node

resetColorInformation

public void resetColorInformation()

Clean all color information in this node and all children.

newColors

boolean newColors()

Returns:

true of the colors of this node were overwritten.

getOwnIndex

public int getOwnIndex()

Returns:

index of this node itself

setParent

public void setParent(DecisionTreeNode parent)

Set parent of this node.

Parameters:

parent - new parent

getOverallColorCount

double getOverallColorCount()

Returns:

overall accumulate count of all color-weights in this node

toString

public final String toString()

Overrides:

toString in class Object

setPrefix

public void setPrefix(String pf)

Set information about this node, e.g. condition this branch needs to fulfill.

Parameters:

pf - string describing condition

getStringSummary

public abstract String getStringSummary()

Returns:

string summary of node content (split, leaf info...)

saveToPredictorParams

public void saveToPredictorParams(ModelContentWO predParams,
                                  boolean saveColorsAndKeys)

Save node to a model content object.

Parameters:

predParams - configuration object to attach decision tree to

saveColorsAndKeys - whether to save the colors and the row keys

saveNodeInternalsToPredParams

public abstract void saveNodeInternalsToPredParams(ModelContentWO pConf,
                                                   boolean saveKeysAndPatterns)

Save internal node settings to a model content object.

Parameters:

pConf - configuration object to attach decision tree to

saveKeysAndPatterns - whether to save the keys and patterns

loadFromPredictorParams

public void loadFromPredictorParams(ModelContentRO predParams)
                             throws InvalidSettingsException

Load node from a model content object.

Parameters:

predParams - configuration object to load decision tree from

Throws:

InvalidSettingsException - if something goes wrong

createNodeFromPredictorParams

public static DecisionTreeNode createNodeFromPredictorParams(ModelContentRO predParams,
                                                             DecisionTreeNode parent)
                                                      throws InvalidSettingsException

Creates a new DecisionTreeNode (and all it's children!) based on an model content object.

Parameters:

predParams - configuration object

parent - the parent node (or null if this is the root)

Returns:

newly created DecisionTreeNode

Throws:

InvalidSettingsException - if something goes wrong

loadNodeInternalsFromPredParams

public abstract void loadNodeInternalsFromPredParams(ModelContentRO pConf)
                                              throws InvalidSettingsException

Load internal node settings from a model content object.

Parameters:

pConf - configuration object to load decision tree from

Throws:

InvalidSettingsException - if something goes wrong

getChildCount

public abstract int getChildCount()

Specified by:

getChildCount in interface TreeNode

Returns:

count of children

getIndex

public abstract int getIndex(TreeNode node)

Returns the index of node in the receivers children. If the receiver does not contain node, -1 will be returned.

Specified by:

getIndex in interface TreeNode

Parameters:

node - that supposedly is a child of this one

Returns:

index of node (or -1 if not found)

getChildAt

public abstract DecisionTreeNode getChildAt(int pos)

Specified by:

getChildAt in interface TreeNode

Parameters:

pos - position of child

Returns:

child node at index

getParent

public final DecisionTreeNode getParent()

Specified by:

getParent in interface TreeNode

Returns:

parent of node

getCountOfSubtree

public abstract int getCountOfSubtree()

Returns the count of the subtree.

Returns:

the count of the subtree

isLeaf

public abstract boolean isLeaf()

Specified by:

isLeaf in interface TreeNode

Returns:

true if node is a leaf

children

public abstract Enumeration<DecisionTreeNode> children()

Specified by:

children in interface TreeNode

Returns:

enumeration of all children

getAllowsChildren

public abstract boolean getAllowsChildren()

Specified by:

getAllowsChildren in interface TreeNode

Returns:

true if the receiver allows children

getPrefix

public String getPrefix()

Returns the prefix of this node representing the condition.

Returns:

the prefix of this node representing the condition

addColorToMap

protected void addColorToMap(Color col,
                             double weight)

Adds the given color to the color map.

Parameters:

col - the color to add

weight - the weight for the color count

getCustomData

public Object getCustomData()

To get the custom data object.

Returns:

the custom data object

setCustomData

public void setCustomData(Object customData)

To set a custom data object.

Parameters:

customData - the custom data object to set

setCoveredColors

public void setCoveredColors(HashMap<Color,Double> coveredColors)

Sets the covered colors distribution directly.

Parameters:

coveredColors - the color distribution to set