/// <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);
}
/// <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);
}
/// <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;
}
/**
* 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;
}
/**
* Generates statistics (used for splitting) based on the
* current position in the tree (as defined by an
* attribute mask).
*
* @return A Vector that contains the available attributes.
* Each attribute's internal statistics array is
* populated with appropriate data. The supplied
* stats array is filled with counts of the number of
* examples that fall into each of the target classes
* at the current position in the tree.
*/
public List<Attribute> generateStats(AttributeMask mask, int[] stats)
{
// First, we fill the stats array - this is not the
// most efficient approach, since we're looping through
// the data several times.
getExampleCounts(mask, DatasetUse.getTrainingExamples(), stats, null);
// Now, we have to go through the attribute mask
// and locate the attributes that are still available.
List<Attribute> results = new List<Attribute>();
// Create a new mask that we can modify.
AttributeMask newMask = new AttributeMask(mask);
// We don't use position 0, that's where the target attribute is.
for (int i = 1; i < mask.getNumAttributes(); i++)
{
if (newMask.isMasked(i) == AttributeMask.UNUSED)
{
// This attribute is available, so we calculate stats for it.
Attribute att = null;
try
{
att = DatasetUse.getAttributeByNum(i);
}
catch (Exception e)
{
// This can't happen!!
return null;
}
int[][] attStats = att.getStatsArray();
// Modify the mask and fill in the arrays.
for (int j = 0; j < att.getNumValues(); j++)
{
newMask.mask(i, j);
getExampleCounts(newMask, DatasetUse.getTrainingExamples(), attStats[j], null);
}
// Reset the mask.
newMask.unmask(i);
results.Add(att);
}
}
return results;
}
/**
* Classifies all examples in the current set of
* examples, by target attribute value. The
* attribute mask determines which examples from the
* dataset form the current example set.
*
* @param mask The current attribute mask that
* indicates which examples from the dataset
* should be considered.
*
* @param conclusion The method expects the parameter
* to be an array of size 8. Positions in
* the array are filled with the following
* values.
*
* <ul>
* <li><i>Position 0</i> - Records the
* index of the most common target attribute
* value from the training dataset.
*
* <li><i>Position 1</i> - Records the number
* of training examples from the dataset that
* reach the current position.
*
* <li><i>Position 2</i> - Records the number
* of training examples from the dataset
* that would be correcly classified
* <i>if a leaf with the most common training
* target classification</i> were added at the
* current position.
*
* <li><i>Position 3</i> - Records the number
* if testing examples from the dataset
* that would be correctly classified
* <i>if a leaf with the most common training
* target classification</i> were added at the
* current position.
*
* <li><i>Position 4</i> - Records the index
* of the most common target attribute
* value from the testing dataset.
*
* <li><i>Position 5</i> - Records the number
* of testing examples from the dataset that
* reach the current position.
*
* <li><i>Position 6</i> - Records the number
* of testing examples from the dataset
* that would be correcly classified
* <i>if a leaf with the most common testing
* target classification</i> were added at the
* current position.
*
* <li><i>Position 7</i> - Records the number
* if training examples from the dataset
* that would be correctly classified
* <i>if a leaf with the most common testing
* target classification</i> were added at the
* current position.
* </ul>
*
* @param trainingCounts The method expects the parameter to be
* an array with a size equal to the number of
* target attribute values. Each position in
* the array is filled with a corresponding count
* of the number of training examples that fall into
* that particular target class, at the current
* position in the tree. This parameter can be null
* if training count data is not required.
*
* @param testingCounts The method expects the parameter to be
* an array with a size equal to the number of
* target attribute values. Each position in the
* array is filled with a corresponding count of
* the number of testing examples that fall into
* that particular target class, at the current
* position in the tree. This parameter can be null
* if testing count data is not required.
*
* @param examples The method expects the parameter to be
* an array with a size equal to the number of
* training examples in the dataset. Each entry in
* the array is marked with true or false, depending
* on whether or not a particular example reaches
* the current position.
*
* @return DATASET_MIXED_CONCL if the examples have
* multiple, different target attribute values.
* DATASET_IDENT_CONCL if all the exmamples have
* the same target attribute value.
* DATASET_EMPTY if there are no examples in the
* current example set.
*
* <p>
* If the result is DATASET_IDENT_CONCL, the
* index of the single target attribute value
* is returned in <code>conclusion[0]</code>. If
* the result is DATASET_EMPTY, the index of the
* most common target attribute value is returned
* in <code>conclusion[0]</code>.
*/
public int classifyExamples(AttributeMask mask,
int[] conclusion,
int[] trainingCounts,
int[] testingCounts,
bool[] examples)
{
if (mask == null || conclusion == null)
throw new Exception("Mask or conclusion array is null.");
// Determine the number of target attribute values
// and create some storage space for our counts.
int[] currTrainingCounts = null;
int[] currTestingCounts = null;
if (trainingCounts != null)
currTrainingCounts = trainingCounts;
else
currTrainingCounts = new
int[DatasetUse.getTargetAttribute().getNumValues()];
if (testingCounts != null)
currTestingCounts = testingCounts;
else
currTestingCounts = new int[currTrainingCounts.Length];
getExampleCounts(mask, DatasetUse.getTrainingExamples(), currTrainingCounts, examples);
getExampleCounts(mask, DatasetUse.getTestingExamples(), currTestingCounts, null);
// Init results.
conclusion[0] = 0; // Training target attribute value index.
conclusion[1] = 0; // Total number of training examples
// reaching node.
conclusion[2] = 0; // Number of training examples correctly
// classified if this node were a leaf
// with most common training target value.
conclusion[3] = 0; // Number of testing examples correctly
// classified if this node were a leaf
// with most common training target value.
conclusion[4] = 0; // Testing target attribute value index.
conclusion[5] = 0; // Total number of testing examples
// reaching node.
conclusion[6] = 0; // Number of testing examples correctly
// classified if this node were a leaf
// with most common testing target value.
conclusion[7] = 0; // Number of training examples correctly
// classified if this node were a leaf
// with most common testing target value.
// Examine the results and determine the conclusion.
int result = DATASET_EMPTY;
for (int i = 0; i < currTrainingCounts.Length; i++)
{
// Increment # of examples that reach this position.
conclusion[1] += currTrainingCounts[i];
conclusion[5] += currTestingCounts[i];
if (result == DATASET_EMPTY && currTrainingCounts[i] != 0)
result = DATASET_IDENT_CONCL;
else if (result == DATASET_IDENT_CONCL && currTrainingCounts[i] != 0)
result = DATASET_MIXED_CONCL;
if (currTrainingCounts[i] >= currTrainingCounts[conclusion[0]])
{
// This target value is more common in the training set.
conclusion[0] = i;
conclusion[2] = currTrainingCounts[i];
conclusion[3] = currTestingCounts[i];
}
if (currTestingCounts[i] >= currTestingCounts[conclusion[4]])
{
// This target value is more common in the testing set.
conclusion[4] = i;
conclusion[6] = currTestingCounts[i];
conclusion[7] = currTrainingCounts[i];
}
}
return result;
}
/**
* Fills the supplied array with the number of examples
* from the current dataset that fall into each of the
* target categories (based on the attribute mask).
*
* @param mask The mask that determines which examples
* reach the current position in the decision
* tree.
*
* @param examples An iteration over a series of
* examples from the current dataset.
*
* @param counts The method expects the parameter to be
* an array with a size equal to the number of
* target attribute values. Each position in
* the array is filled with a corresponding count
* of the number of training examples that fall into
* that particular target class, at the current
* position in the decision tree.
*
* @param reachedHere The method expects the parameter
* to be an array with a size equal to the
* <i>total</i> number of examples being examined.
* Each cell in the array is set to true or
* false, depending on whether or not the
* corresponding example reaches the current
* position in the decision tree.
*/
private void getExampleCounts(AttributeMask mask, IEnumerator<int[]> examples, int[] counts, bool[] reachedHere)
{
// Zero any values currently in stats.
for (int i = 0; i < counts.Length; i++)
{
counts[i] = 0;
}
int j = 0;
// Loop through and get totals.
while (examples.MoveNext())
{
int[] example = (int[])examples.Current;
if (mask.matchMask(example))
{
counts[example[0]]++; // Increment appropriate
// target count.
if (reachedHere != null && reachedHere.Length > j)
reachedHere[j] = true;
}
j++;
}
}
/// <summary>
/// Builds a new mask, copying the state of the
/// supplied mask to the new mask.
/// </summary>
/// <param name="mask">An Attribute mask</param>
public AttributeMask(AttributeMask mask)
{
Mask = (int[])mask.Mask.Clone();
}
/// <summary>
/// Builds a new mask, copying the state of the
/// supplied mask to the new mask.
/// </summary>
/// <param name="mask">An Attribute mask</param>
public AttributeMask(AttributeMask mask)
{
Mask = (int[])mask.Mask.Clone();
}
/// <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;
}
