/// <summary>
/// Attaches a new internal node to the supplied node,
/// along the specified arc.
/// </summary>
/// <param name="parent">The node in the current tree to attach
/// the internal node to. If the node is null, the
/// new internal node becomes the root of the tree.</param>
/// <param name="arcNum">The arc number (or attribute value
/// index) along which to attach the new node.</param>
/// <param name="attributePosition">The position of the
/// attribute used to split at the new node, relative
/// to the other attributes in the dataset.</param>
/// <param name="att">The attribute used to split at the new
/// node.</param>
/// <returns>A reference to the new internal node.</returns>
public DecisionTreeNode addInternalNode(DecisionTreeNode parent,
int arcNum,
int attributePosition,
Attribute att,
AttributeMask mask,
int numTrainingExamplesReachHere,
int bestTrainingTargetIndex,
int numTrainingEgsCorrectClassUsingBestTrainingIndex,
int numTestingEgsCorrectClassUsingBestTrainingIndex,
int numTestingExamplesReachHere,
int bestTestingTargetIndex,
int numTestingEgsCorrectClassUsingBestTestingIndex,
int numTrainingEgsCorrectClassUsingBestTestingIndex)
{
// Create a new internal node.
DecisionTreeNode internalNode = new DecisionTreeNode(parent, att.getName(), att.getValueNames(), mask);
// Set the node statistics.
internalNode.setTrainingStats(numTrainingExamplesReachHere,
bestTrainingTargetIndex,
numTrainingEgsCorrectClassUsingBestTrainingIndex,
numTestingEgsCorrectClassUsingBestTrainingIndex);
internalNode.setTestingStats(numTestingExamplesReachHere,
bestTestingTargetIndex,
numTestingEgsCorrectClassUsingBestTestingIndex,
numTrainingEgsCorrectClassUsingBestTestingIndex);
// Update the tree statistics.
InternalNodes++;
// Now, attach the new internal node to the supplied node.
if (parent != null)
parent.setChild(arcNum, internalNode);
// Add a reference to the new node to the node list.
Nodes.Add(internalNode);
return internalNode;
}
private int _trainTrainingCorrectClass; // Number of training examples
#endregion Fields
#region Constructors
public DecisionTreeNode(DecisionTreeNode parent,
String label,
Object[] arcLabels,
AttributeMask mask)
{
if (label == null || mask == null)
throw new Exception("Label or attribute mask is null.");
Parent = parent;
NodeLabel = label;
ArcLabels = arcLabels;
AttMask = mask;
// The number of possible children for the node is
// equal to the number of values for this attribute.
// If the arc label array was null, then this is
// a leaf node that has no children.
if (arcLabels != null)
Children = new DecisionTreeNode[ArcLabels.Length];
// The node is initially unflagged.
NodeFlag = -2;
}
/**
* Attaches the supplied node at specified child position.
* The method does not check to see if a node already
* exists at the insertion position.
* If this node is a leaf node, the method has no effect.
*
* @throws IndexOutOfBoundsException If the supplied
* child number is less than 0 or greater
* than the number of possible children (which
* is equivalent to the number of attribute
* values) at this node.
*/
public void setChild(int childNum, DecisionTreeNode node)
{
if (isLeaf()) return;
if (childNum < 0 || childNum > (Children.Length - 1))
throw new Exception("Cannot add child at position " + childNum + ".");
Children[childNum] = node;
}
/**
* Searches through the current node's list of children
* and attempts to locate a child that matches the
* supplied child. If a match is found, the reference
* to the child is removed, leaving a vacant arc.
*/
public void removeChild(DecisionTreeNode child)
{
// Search through the list of children, looking for a match.
for (int i = 0; i < Children.Length; i++)
{
if (Children[i] == child)
{
Children[i] = null;
return;
}
}
}
/**
* Returns the position of the supplied child beneath
* the parent node. If this node is a leaf, or the
* supplied node is not a child of this node, the method
* returns -1.
*
* @param child The 'potential' child of this node.
*
* @return The position of the supplied child beneath
* this node, or -1 if this is a leaf node
* or the node is not a parent of the
* child.
*/
public int getChildPosition(DecisionTreeNode child)
{
if (this.isLeaf()) return -1;
for (int i = 0; i < Children.Length; i++)
if (Children[i] == child) return i;
return -1;
}
/**
* An implementation of the recursive decision tree
* pessimistic pruning algorithm. Given a parent
* node, the method will prune all the branches
* below the node.
*
* @param node The root node of the tree to prune.
*
* @param error A <code>double</code> array of size 1. The
* array is used to store the current error value.
*
* @return <code>true</code> if an entire subtree was successfully
* pruned, or <code>false</code> otherwise.
*/
public bool prunePessimisticDT(DecisionTreeNode node, double[] error)
{
// Post-order walk through the tree, marking
// our path as we go along.
if (node.isLeaf())
{
if (node.getTrainingEgsAtNode() == 0)
{
error[0] = 0;
return true;
}
else
{
// We do the error calculation in two steps -
// Here we multiply the error value by the number
// of examples that reach the node. When the method
// is called recursively, this value will be divided
// by the number of examples that reach the parent
// node (thus weighting the error from each child).
int errors1 = (int)node.getTrainingEgsAtNode() - node.getTrainingEgsCorrectClassUsingBestTrainingIndex();
double p1 = (double)(errors1 + 1.0) / (node.getTrainingEgsAtNode() + 2.0);
error[0] = node.getTrainingEgsAtNode() * errorBar(p1, node.getTrainingEgsAtNode()) + errors1;
return true;
}
}
// We're at an internal node, so compute the error
// of the children and use the result to determine
// if we prune or not.
double errorSum = 0;
for (int i = 0; i < node.getArcLabelCount(); i++)
{
// Mark our current path.
Tree.flagNode(node, i);
if (!prunePessimisticDT(node.getChild(i), error))
{
Tree.flagNode(node, -2);
return false;
}
errorSum += error[0];
}
// Mark the node as our current target.
Tree.flagNode(node, -1);
// Get the worst-case performance of this node.
double errorWorst;
if (node.getTrainingEgsAtNode() == 0)
{
error[0] = 0;
return true;
}
int errors = (int)node.getTrainingEgsAtNode() - node.getTrainingEgsCorrectClassUsingBestTrainingIndex();
double p = (double)(errors + 1.0) / (node.getTrainingEgsAtNode() + 2.0);
errorWorst = (double)node.getTrainingEgsAtNode() * errorBar(p, node.getTrainingEgsAtNode()) + errors;
DecisionTreeNode newNode = node;
if (errorWorst < errorSum)
{
// We need to "prune" this node to a leaf.
DecisionTreeNode parent = node.getParent();
int arcNum = -1;
if (parent != null)
{
arcNum = parent.getChildPosition(node);
}
Tree.pruneSubtree(node);
// Figure out the label for the new leaf.
String label = null;
try
{
label = DatasetUse.getTargetAttribute().getAttributeValueByNum(node.getTrainingBestTarget());
}
catch (Exception e)
{
// Should never happen.
//e.printStackTrace();
}
node.getMask().mask(0, node.getTrainingBestTarget());
newNode =
Tree.addLeafNode(parent, arcNum, label,
node.getMask(),
node.getTrainingEgsAtNode(),
node.getTrainingBestTarget(),
node.getTrainingEgsCorrectClassUsingBestTrainingIndex(),
node.getTestingEgsCorrectClassUsingBestTrainingIndex(),
node.getTestingEgsAtNode(),
node.getTestingBestTarget(),
node.getTestingEgsCorrectClassUsingBestTestingIndex(),
node.getTrainingEgsCorrectClassUsingBestTestingIndex());
}
// Update the count.
if (newNode.isLeaf())
{
error[0] = errorWorst;
}
else
{
error[0] = errorSum;
}
// All finished, unmark the node if it still exists.
Tree.flagNode(node, -2);
return true;
}
/**
* An implementation of the recursive decision tree
* learning algorithm. Given a parent node and an arc
* number, the method will attach a new decision 'sub'-tree
* below the parent node.
*
* @param parent The parent node for the new decision tree.
*
* @param arcNum The arc number (or path) along which the
* new subtree will be attached.
*
* @return true if an entire subtree was successfully added,
* false otherwise.
*/
public bool learnDT(DecisionTreeNode parent, int arcNum)
{
AttributeMask mask;
if (parent == null)
{
// We have to add at the root.
mask = new AttributeMask(DatasetUse.getNumAttributes());
}
else
{
mask = new AttributeMask(parent.getMask());
// Mask off the specified arc number.
try
{
mask.mask(DatasetUse.getAttributePosition(parent.getLabel()), arcNum);
}
catch (Exception e)
{
//e.printStackTrace();
return false;
}
}
// Now, classify the examples at the current position.
int[] conclusion = new int[8];
int result = classifyExamples(mask, conclusion, null, null, null);
Attribute target = DatasetUse.getTargetAttribute();
int numTargetVals = target.getNumValues();
String label;
if (result == DATASET_EMPTY)
{
// If no examples reach our current position
// we add a leaf with the most common target
// classfication for the parent node.
// Save testing results.
int numTestingExamplesReachHere = conclusion[5];
int bestTestingTargetIndex = conclusion[4];
int numTestingExamplesCorrectClass = conclusion[6];
int numTrainingExamplesCorrectClass = conclusion[7];
classifyExamples(parent.getMask(), conclusion, null, null, null);
try
{
label = target.getAttributeValueByNum(conclusion[0]);
}
catch (Exception e)
{
return false;
}
// We have to grab the counts again for the testing data...
int[] currTestingCounts = new int[target.getNumValues()];
getExampleCounts(mask, DatasetUse.getTestingExamples(), currTestingCounts, null);
// Mask target value and add a leaf to the tree.
mask.mask(0, conclusion[0]);
DecisionTreeNode node = Tree.addLeafNode(parent,
arcNum,
label,
mask,
0,
conclusion[0],
0,
currTestingCounts[conclusion[0]],
numTestingExamplesReachHere,
bestTestingTargetIndex,
numTestingExamplesCorrectClass,
numTrainingExamplesCorrectClass);
return true;
}
if (result == DATASET_IDENT_CONCL)
{
// Pure result - we can add a leaf node with the
// correct target attribute value.
try
{
label = target.getAttributeValueByNum(conclusion[0]);
}
catch (Exception e)
{
//e.printStackTrace();
return false;
}
// Mask target value and add a leaf to the tree.
mask.mask(0, conclusion[0]);
DecisionTreeNode node = Tree.addLeafNode(parent,
arcNum,
label,
mask,
conclusion[1],
conclusion[0],
conclusion[2],
conclusion[3],
conclusion[5],
conclusion[4],
conclusion[6],
conclusion[7]);
return true;
}
// Mixed conclusion - so we have to select
// an attribute to split on, and then build a
// new internal node with that attribute.
// First, generate statistics - this may take awhile.
int[] nodeStats = new int[numTargetVals];
List<Attribute> availableAtts = generateStats(mask, nodeStats);
if (availableAtts.Count == 0)
{
// No attributes left to split on - so use
// the most common target value at the current position.
try
{
label = target.getAttributeValueByNum(conclusion[0]);
}
catch (Exception e)
{
//e.printStackTrace();
return false;
}
mask.mask(0, conclusion[0]);
DecisionTreeNode node = Tree.addLeafNode(parent,
arcNum,
label,
mask,
conclusion[1],
conclusion[0],
conclusion[2],
conclusion[3],
conclusion[5],
conclusion[4],
conclusion[6],
conclusion[7]);
return true;
}
// Choose an attribute, based on the set of
// available attributes.
List<double> results = new List<double>();
Attribute att = chooseAttribute(availableAtts, nodeStats, results);
int attPos;
try
{
attPos = DatasetUse.getAttributePosition(att.getName());
}
catch (Exception e)
{
//e.printStackTrace();
return false;
}
DecisionTreeNode newParent = Tree.addInternalNode(parent,
arcNum,
attPos,
att,
mask,
conclusion[1],
conclusion[0],
conclusion[2],
conclusion[3],
conclusion[5],
conclusion[4],
conclusion[6],
conclusion[7]);
// Now, recursively decend along each branch of the new node.
for (int j = 0; j < newParent.getArcLabelCount(); j++)
{
// Recursive call.
if (!learnDT(newParent, j)) return false;
}
return true;
}
public void ExtractRulesFromDTTree(DecisionTreeNode currentRoot)
{
if (currentRoot.isLeaf())
{
BackTrackRule(currentRoot);
}
else
{
DecisionTreeNode temp = currentRoot;
for (int i = 0; i < temp.Children.Length; i++)
{
currentRoot = temp.Children[i];
ExtractRulesFromDTTree(currentRoot);
}
}
}
private void BackTrackRule(DecisionTreeNode currentNode)
{
string rule;
string[] Temp = {"IF", "AND","THEN"};
DecisionTreeNode tempParentNode = currentNode;
DecisionTreeNode tempCurrentNode = currentNode;
Rule newRule = new Rule();
// Xây dựng luật
rule = Temp[2] + " \"" + DatasetUse.Attributes[0].Name + "\" = " + "\'"+ currentNode.NodeLabel +"\'";
newRule.AddAttributeTarget(DatasetUse.Attributes[0].Name, currentNode.NodeLabel);
while (tempCurrentNode.Parent != null)
{
tempParentNode = tempCurrentNode.Parent;
// Kiểm tra đúng cha con
int i;
for( i = 0 ; i < tempParentNode.Children.Length; i++)
{
if(tempParentNode.Children[i] == tempCurrentNode)
{
break;
}
}
rule = Temp[1] + " \"" + tempParentNode.NodeLabel
+ "\" = \'" + tempParentNode.ArcLabels[i] + "\' " + rule;
newRule.AddAttributeCondition(tempParentNode.NodeLabel , tempParentNode.ArcLabels[i].ToString());
tempCurrentNode = tempParentNode;
}
rule = rule.Substring(rule.IndexOf(Temp[1]) + Temp[1].Length);
rule = Temp[0] + rule;
newRule.SetRule(rule);
newRule.AttributeConditions.Reverse();
newRule.AttributeConditionValues.Reverse();
newRule.AttributeTargets.Reverse();
newRule.AttributeTargetValues.Reverse();
// Thêm một luật mới vào danh sách các luật rút ra
ListRules.ListRules.Add(newRule);
}
/**
* An implementation of the recursive decision tree
* reduced error pruning algorithm. Given a parent
* node, the method will prune all the branches
* below the node.
*
* @param node The root node of the tree to prune.
*
* @param error A <code>double</code> array of size 1. The
* array is used to store the current error value.
*
* @return <code>true</code> if an entire subtree was successfully
* pruned, or <code>false</code> otherwise.
*/
public bool pruneReducedErrorDT(DecisionTreeNode node, double[] error)
{
if (node.isLeaf())
{
error[0] = node.getTestingEgsAtNode() - node.getTestingEgsCorrectClassUsingBestTrainingIndex();
return true;
}
// We're at an internal node, so compute the error
// of the children and use the result to determine
// if we prune or not.
double errorSum = 0;
for (int i = 0; i < node.getArcLabelCount(); i++)
{
// Mark our current path.
Tree.flagNode(node, i);
if (!pruneReducedErrorDT(node.getChild(i), error))
{
Tree.flagNode(node, -2);
return false;
}
errorSum += error[0];
}
// Mark the node as our current target.
Tree.flagNode(node, -1);
// Get the best-case performance of this node.
double errorBest = node.getTestingEgsAtNode() - node.getTestingEgsCorrectClassUsingBestTestingIndex();
DecisionTreeNode newNode = node;
if (errorBest < errorSum)
{
// We need to "prune" this node to a leaf.
DecisionTreeNode parent = node.getParent();
int arcNum = -1;
if (parent != null)
{
arcNum = parent.getChildPosition(node);
}
Tree.pruneSubtree(node);
// Figure out the label for the new leaf.
String label = null;
try
{
label = DatasetUse.getTargetAttribute().getAttributeValueByNum(node.getTestingBestTarget());
}
catch (Exception e)
{
// Should never happen.
//e.printStackTrace();
}
node.getMask().mask(0, node.getTestingBestTarget());
newNode = Tree.addLeafNode(parent, arcNum, label,
node.getMask(),
node.getTrainingEgsAtNode(),
node.getTestingBestTarget(),
node.getTrainingEgsCorrectClassUsingBestTestingIndex(),
node.getTestingEgsCorrectClassUsingBestTestingIndex(),
node.getTestingEgsAtNode(),
node.getTestingBestTarget(),
node.getTestingEgsCorrectClassUsingBestTestingIndex(),
node.getTrainingEgsCorrectClassUsingBestTestingIndex());
}
// Update the count.
if (newNode.isLeaf())
{
error[0] = errorBest;
}
else
{
error[0] = errorSum;
}
// All finished, unmark the node if it still exists.
Tree.flagNode(node, -2);
return true;
}
/// <summary>
/// Sets the given node's flag value. This method exist
/// here in order to allow the tree to notify
/// TreeChangeListeners that may be tracking the state of
/// the tree.
public void flagNode(DecisionTreeNode node, int flagValue)
{
node.setFlag(flagValue);
}
/**
* Recursively descends through the tree, removing
* the supplied node and any descendants from
* the internal node list.
*
* @param node The root node of the subtree to remove.
*/
public void recursiveRemoveSubtree(DecisionTreeNode node)
{
if (node == null) return;
// First, recursively remove all the node's children.
// (This loop doesn't run if the node is a leaf node)
for (int i = 0; i < node.getArcLabelCount(); i++)
if (node.getChild(i) != null)
recursiveRemoveSubtree(node.getChild(i));
// Remove this node from the vector.
Nodes.Remove(node);
// If the node was a leaf, then we have to update
// the classification statistics.
if (node.isLeaf())
{
TrainingCorrect -=
node.getTrainingEgsCorrectClassUsingBestTrainingIndex();
TestingCorrect -=
node.getTestingEgsCorrectClassUsingBestTrainingIndex();
}
else
InternalNodes--;
}
/**
* Prunes off the subtree starting at the supplied root.
*
* @param pruneRoot The root of the subtree to remove.
*/
public void pruneSubtree(DecisionTreeNode pruneRoot)
{
if (pruneRoot == null)
{
return;
}
// Detach the root node of the subtree.
pruneRoot.detach();
// Once a node has been removed, the tree is no longer complete.
Complete = false;
// Now, tell the tree to remove all the
// node's children.
recursiveRemoveSubtree(pruneRoot);
}
/// <summary>
/// Sets the given node's flag value. This method exist
/// here in order to allow the tree to notify
/// TreeChangeListeners that may be tracking the state of
/// the tree.
public void flagNode(DecisionTreeNode node, int flagValue)
{
node.setFlag(flagValue);
}
/// <summary>
/// Search through the current tree and return the first
/// node without a complete set of children. The arc
/// number for the first missing child is returned in
/// position 0 of the arcNum array.
/// </summary>
/// <param name="node">The node at which to begin the search.</param>
/// <param name="arcNum">An integer array of size 1. The arc
/// number for the first missing child is
/// returned in arcNum[0].</param>
/// <returns>A reference to the first incomplete node
/// in the tree (farthest left). The method
/// returns null if the tree is already complete,
/// or if the tree is empty.</returns>
public DecisionTreeNode findIncompleteNode(DecisionTreeNode node, int[] arcNum)
{
// The search is recursive - at some point, we
// may want to change this to a non-recursive
// algorithm (if we start dealing with extremely
// large trees?)
// Base cases.
if (node == null || node.isLeaf())
{
return null;
}
// Recursive case. This node is not a leaf - so descend.
for (int i = 0; i < node.getArcLabelCount(); i++)
{
DecisionTreeNode nextNode;
if ((nextNode = findIncompleteNode(node.getChild(i), arcNum)) != null)
{
return nextNode;
}
}
if ((arcNum[0] = node.getFirstMissingChild()) >= 0)
{
// We found a node with a missing child, which
// is already stored in arcNum - so return this
// node.
return node;
}
// We searched all the subtrees attached to this
// node, and didn't find anything, so return null.
return null;
}
/// <summary>
/// Attaches a new leaf node to the supplied node, along
/// the specified arc.
/// </summary>
public DecisionTreeNode addLeafNode(DecisionTreeNode parent,
int arcNum,
String label,
AttributeMask mask,
int numTrainingExamplesReachHere,
int bestTrainingTargetIndex,
int numTrainingEgsCorrectClassUsingBestTrainingIndex,
int numTestingEgsCorrectClassUsingBestTrainingIndex,
int numTestingExamplesReachHere,
int bestTestingTargetIndex,
int numTestingEgsCorrectClassUsingBestTestingIndex,
int numTrainingEgsCorrectClassUsingBestTestingIndex)
{
// Create new leaf node.
DecisionTreeNode leaf =
new DecisionTreeNode(parent, label, null, mask);
// Set the node statistics.
leaf.setTrainingStats(numTrainingExamplesReachHere,
bestTrainingTargetIndex,
numTrainingEgsCorrectClassUsingBestTrainingIndex,
numTestingEgsCorrectClassUsingBestTrainingIndex);
leaf.setTestingStats(numTestingExamplesReachHere,
bestTestingTargetIndex,
numTestingEgsCorrectClassUsingBestTestingIndex,
numTrainingEgsCorrectClassUsingBestTestingIndex);
// Update the tree statistics.
TrainingCorrect += numTrainingEgsCorrectClassUsingBestTrainingIndex;
TestingCorrect += numTestingEgsCorrectClassUsingBestTrainingIndex;
// Now, attach the new leaf to the supplied node.
if (parent != null)
parent.setChild(arcNum, leaf);
// Add a reference to the new node to the node list.
Nodes.Add(leaf);
// Determine if the tree is complete.
if (findIncompleteNode((DecisionTreeNode)Nodes[0], new int[1]) == null)
{
Complete = true;
}
return leaf;
}
