GestureRecognitionToolkit  Version: 1.0 Revision: 04-03-15
The Gesture Recognition Toolkit (GRT) is a cross-platform, open-source, c++ machine learning library for real-time gesture recognition.
GRT::Tree Class Reference
Inheritance diagram for GRT::Tree:
GRT::GRTBase GRT::ClusterTree GRT::DecisionTree GRT::RegressionTree

Public Types

enum  TrainingMode { BEST_ITERATIVE_SPILT =0, BEST_RANDOM_SPLIT, NUM_TRAINING_MODES }
 

Public Member Functions

 Tree (const UINT numSplittingSteps=100, const UINT minNumSamplesPerNode=5, const UINT maxDepth=10, const bool removeFeaturesAtEachSpilt=false, const UINT trainingMode=BEST_ITERATIVE_SPILT)
 
virtual ~Tree (void)
 
virtual bool clear ()
 
virtual bool print () const
 
virtual bool getModel (ostream &stream) const
 
virtual NodedeepCopyTree () const
 
const NodegetTree () const
 
UINT getTrainingMode () const
 
UINT getNumSplittingSteps () const
 
UINT getMinNumSamplesPerNode () const
 
UINT getMaxDepth () const
 
UINT getPredictedNodeID () const
 
bool getRemoveFeaturesAtEachSpilt () const
 
bool setTrainingMode (const UINT trainingMode)
 
bool setNumSplittingSteps (const UINT numSplittingSteps)
 
bool setMinNumSamplesPerNode (const UINT minNumSamplesPerNode)
 
bool setMaxDepth (const UINT maxDepth)
 
bool setRemoveFeaturesAtEachSpilt (const bool removeFeaturesAtEachSpilt)
 
- Public Member Functions inherited from GRT::GRTBase
 GRTBase (void)
 
virtual ~GRTBase (void)
 
bool copyGRTBaseVariables (const GRTBase *GRTBase)
 
string getClassType () const
 
string getLastWarningMessage () const
 
string getLastErrorMessage () const
 
string getLastInfoMessage () const
 
GRTBasegetGRTBasePointer ()
 
const GRTBasegetGRTBasePointer () const
 

Protected Attributes

UINT trainingMode
 
UINT numSplittingSteps
 
UINT minNumSamplesPerNode
 
UINT maxDepth
 
bool removeFeaturesAtEachSpilt
 
Nodetree
 
- Protected Attributes inherited from GRT::GRTBase
string classType
 
DebugLog debugLog
 
ErrorLog errorLog
 
InfoLog infoLog
 
TrainingLog trainingLog
 
TestingLog testingLog
 
WarningLog warningLog
 

Additional Inherited Members

- Static Public Member Functions inherited from GRT::GRTBase
static string getGRTVersion (bool returnRevision=true)
 
static string getGRTRevison ()
 
- Protected Member Functions inherited from GRT::GRTBase
double SQR (const double &x) const
 

Detailed Description

Definition at line 38 of file Tree.h.

Constructor & Destructor Documentation

GRT::Tree::Tree ( const UINT  numSplittingSteps = 100,
const UINT  minNumSamplesPerNode = 5,
const UINT  maxDepth = 10,
const bool  removeFeaturesAtEachSpilt = false,
const UINT  trainingMode = BEST_ITERATIVE_SPILT 
)

Default Constructor

Parameters
UINTnumSplittingSteps: sets the number of steps that will be used to search for the best spliting value for each node. Default value = 100
UINTminNumSamplesPerNode: sets the minimum number of samples that are allowed per node, if the number of samples is below that, the node will become a leafNode. Default value = 5
UINTmaxDepth: sets the maximum depth of the tree. Default value = 10
boolremoveFeaturesAtEachSpilt: sets if a feature is removed at each spilt so it can not be used again. Default value = false
UINTtrainingMode: sets the training mode, this should be one of the TrainingMode enums. Default value = BEST_ITERATIVE_SPILT

Definition at line 25 of file Tree.cpp.

GRT::Tree::~Tree ( void  )
virtual

Default Destructor

Definition at line 40 of file Tree.cpp.

Member Function Documentation

bool GRT::Tree::clear ( )
virtual

This overrides the clear function in the Regressifier base class. It will completely clear the ML module, removing any trained model and setting all the base variables to their default values.

Returns
returns true if the module was cleared succesfully, false otherwise

Reimplemented in GRT::DecisionTree, GRT::ClusterTree, and GRT::RegressionTree.

Definition at line 45 of file Tree.cpp.

Node * GRT::Tree::deepCopyTree ( ) const
virtual

Deep copies the tree, returning a pointer to the new tree. The user is in charge of cleaning up the memory so must delete the pointer when they no longer need it. NULL will be returned if the tree could not be copied.

Returns
returns a pointer to a deep copy of the tree

Reimplemented in GRT::DecisionTree, GRT::ClusterTree, and GRT::RegressionTree.

Definition at line 78 of file Tree.cpp.

UINT GRT::Tree::getMaxDepth ( ) const

Gets the maximum depth of the tree.

Returns
returns the maximum depth of the tree

Definition at line 103 of file Tree.cpp.

UINT GRT::Tree::getMinNumSamplesPerNode ( ) const

Gets the minimum number of samples that are allowed per node, if the number of samples at a node is below this value then the node will automatically become a leaf node.

Returns
returns the minimum number of samples that are allowed per node

Definition at line 99 of file Tree.cpp.

bool GRT::Tree::getModel ( ostream &  stream) const
virtual

This function adds the current model to the formatted stream. This function should be overwritten by the derived class.

Parameters
ostream&file: a reference to the stream the model will be added to
Returns
returns true if the model was added successfully, false otherwise

Reimplemented in GRT::DecisionTree.

Definition at line 67 of file Tree.cpp.

UINT GRT::Tree::getNumSplittingSteps ( ) const

Gets the number of steps that will be used to search for the best spliting value for each node.

If the trainingMode is set to BEST_ITERATIVE_SPILT, then the numSplittingSteps controls how many iterative steps there will be per feature. If the trainingMode is set to BEST_RANDOM_SPLIT, then the numSplittingSteps controls how many random searches there will be per feature.

Returns
returns the number of steps that will be used to search for the best spliting value for each node

Definition at line 95 of file Tree.cpp.

UINT GRT::Tree::getPredictedNodeID ( ) const

This function returns the predictedNodeID, this is ID of the leaf node that was reached during the last prediction call

Returns
returns the predictedNodeID, this will be zero if the tree does not exist or predict has not been called

Definition at line 107 of file Tree.cpp.

bool GRT::Tree::getRemoveFeaturesAtEachSpilt ( ) const

Gets if a feature is removed at each spilt so it can not be used again.

Returns
returns true if a feature is removed at each spilt so it can not be used again, false otherwise

Definition at line 116 of file Tree.cpp.

UINT GRT::Tree::getTrainingMode ( ) const

Gets the current training mode. This will be one of the TrainingModes enums.

Returns
returns the training mode

Definition at line 91 of file Tree.cpp.

const Node * GRT::Tree::getTree ( ) const

Gets a pointer to the root node of the tree. NULL will be returned if the tree model has not be trained.

Returns
returns a const pointer to the tree

Definition at line 87 of file Tree.cpp.

bool GRT::Tree::print ( ) const
virtual

Prints the tree to std::cout.

Returns
returns true if the model was printed

Reimplemented in GRT::ClusterTree, and GRT::RegressionTree.

Definition at line 56 of file Tree.cpp.

bool GRT::Tree::setMaxDepth ( const UINT  maxDepth)

Sets the maximum depth of the tree, any node that reaches this depth will automatically become a leaf node. Value must be larger than zero.

Parameters
UINTmaxDepth: the maximum depth of the tree
Returns
returns true if the parameter was set, false otherwise

Definition at line 147 of file Tree.cpp.

bool GRT::Tree::setMinNumSamplesPerNode ( const UINT  minNumSamplesPerNode)

Sets the minimum number of samples that are allowed per node, if the number of samples at a node is below this value then the node will automatically become a leaf node. Value must be larger than zero.

Parameters
UINTminNumSamplesPerNode: the minimum number of samples that are allowed per node
Returns
returns true if the parameter was set, false otherwise

Definition at line 138 of file Tree.cpp.

bool GRT::Tree::setNumSplittingSteps ( const UINT  numSplittingSteps)

Sets the number of steps that will be used to search for the best spliting value for each node.

If the trainingMode is set to BEST_ITERATIVE_SPILT, then the numSplittingSteps controls how many iterative steps there will be per feature. If the trainingMode is set to BEST_RANDOM_SPLIT, then the numSplittingSteps controls how many random searches there will be per feature.

A higher value will increase the chances of building a better model, but will take longer to train the model. Value must be larger than zero.

Parameters
UINTnumSplittingSteps: sets the number of steps that will be used to search for the best spliting value for each node.
Returns
returns true if the parameter was set, false otherwise

Definition at line 129 of file Tree.cpp.

bool GRT::Tree::setRemoveFeaturesAtEachSpilt ( const bool  removeFeaturesAtEachSpilt)

Sets if a feature is removed at each spilt so it can not be used again. If true then the best feature selected at each node will be removed so it can not be used in any children of that node. If false, then the feature that provides the best spilt at each node will be used, regardless of how many times it has been used again.

Parameters
boolremoveFeaturesAtEachSpilt: if true, then each feature is removed at each spilt so it can not be used again
Returns
returns true if the parameter was set, false otherwise

Definition at line 156 of file Tree.cpp.

bool GRT::Tree::setTrainingMode ( const UINT  trainingMode)

Sets the training mode, this should be one of the TrainingModes enums.

Parameters
constUINT trainingMode: the new trainingMode, this should be one of the TrainingModes enums
Returns
returns true if the trainingMode was set successfully, false otherwise

Definition at line 120 of file Tree.cpp.


The documentation for this class was generated from the following files: