KSquare Utilities
Classes | Typedefs | Enumerations | Functions | Variables
KKMLL Namespace Reference

Namespace for all K^2 Machine Learning code. More...

Classes

class  Attribute
 describes a single Feature, Type and possible values. More...
 
class  AttributeList
 
class  AttributeTypeVector
 
class  BinaryClassParms
 Similar to SVMparam except it is specialized for two classes. More...
 
class  BinaryClassParmsList
 
class  ClassAssignments
 Binds MLClass objects to the appropriate number that the Learning Algorithm expects. More...
 
class  ClassificationBiasMatrix
 Assists in adjusting a Classifiers output for bias of a classifier. More...
 
class  Classifier2
 
class  Classifier2List
 
class  ClassProb
 Used to record probability for a specified class; and a list of classes. More...
 
class  ClassProbList
 
class  ClassStatistic
 Used by routines that retrieve Class statistics from FeatureVectorList instances. More...
 
class  ClassStatisticList
 
class  ConfusionMatrix2
 A confusion matrix object that is used to record the results from a CrossValidation. <see also cref="CrossValidation" More...
 
class  ConfussionMatrix2List
 
class  CrossValidation
 A class that is meant to manage a n-Fold Cross Validation. More...
 
class  CrossValidationMxN
 
class  CrossValidationVoting
 
class  DuplicateImage
 
class  DuplicateImageList
 
class  DuplicateImages
 Detects duplicate images in a given FeaureVectorList objects. More...
 
class  ExtractExampleFileName
 
class  ExtractFeatureData
 
class  FactoryFVProducer
 Responsible for creating a FeatureFectorProducer instance. More...
 
class  FeatureEncoder
 
class  FeatureEncoder2
 
class  FeatureFileIO
 Base class for all FeatureFileIO classes. More...
 
class  FeatureFileIOArff
 Support the writing of ARFF Formatted Feature Files. More...
 
class  FeatureFileIOC45
 Supports the reading and writing of feature data from C45 formated feature files. More...
 
class  FeatureFileIOColumn
 Supports a simple Feature File format where each column represents a example and each row a feature value. More...
 
class  FeatureFileIODstWeb
 
class  FeatureFileIORoberts
 Supports the writing of Feature Data to a file that can then be read by OpenDT. More...
 
class  FeatureFileIOSparse
 Supports the reading and writing of Sparse feature files similar to the ones libSVM use. More...
 
class  FeatureFileIOUCI
 Supports the reading and writing of Feature data from a file format commonly used by many dataset's in the UCI repository. More...
 
class  FeatureNumList
 Keeps track of selected features. More...
 
class  FeatureVector
 Represents a Feature Vector of a single example, labeled or unlabeled. More...
 
class  FeatureVectorList
 Container class for FeatureVector derived objects. More...
 
class  FeatureVectorProducer
 A abstract class that is meant to compute a FeatureVector from a source image. More...
 
class  FileDesc
 Provides a detailed description of the attributes of a dataset. More...
 
class  FileDescList
 Container class file 'FileDesc' instances. More...
 
class  GrayScaleImagesFV
 Specialized version of KKMLL::FeatureVector that will be used to represent the features of a Shrimp. More...
 
class  GrayScaleImagesFVList
 
class  GrayScaleImagesFVProducer
 
class  GrayScaleImagesFVProducerFactory
 
class  ImageDataTreeEntry
 
class  ImageFeaturesDataIndexed
 
class  ImageFeaturesNameIndexed
 
class  ImageFeaturesNodeKey
 
class  KKMLVariables
 
class  MLClass
 Represents a "Class" in the Machine Learning Sense. More...
 
class  MLClassIndexList
 Maintains a list of classes and their associated integer index. More...
 
class  MLClassList
 Maintains a list of MLClass instances. More...
 
class  Model
 Base class to all Learning Algorithms. More...
 
class  ModelDual
 Will implement the Dual Classifier Model. More...
 
class  ModelKnn
 
class  ModelOldSVM
 A specialization of 'Model'; meant to Wrap the original version of 'SvmModel' class. It will allow us to use the original implementation using version 2.39 of LibSVM. More...
 
class  ModelParam
 Abstract Base class for Machine Learning parameters. More...
 
class  ModelParamDual
 
class  ModelParamKnn
 
class  ModelParamOldSVM
 
class  ModelParamSvmBase
 
class  ModelParamUsfCasCor
 This class encapsulates are the information necessary to build a UsfCasCor class. More...
 
class  ModelSvmBase
 
class  ModelUsfCasCor
 
class  NormalizationParms
 Normalization Parameters; calculation and implementation. More...
 
class  Orderings
 Used to maintain multiple orderings of a single list of FeatureVector objects. More...
 
class  ProbNamePair
 
class  SizeDistribution
 Used to keep track of examples by size; typically used by 'CrossValidation'; for each example predicted it would call the ;Increment' method keeping track. More...
 
class  SVMModel
 
class  SVMparam
 This class encapsulates are the information necessary to build a SVMModel class. More...
 
class  TrainingClass
 Specify where training examples and other related data for a MLClass that is needed to train a classifier. More...
 
class  TrainingClassList
 
class  TrainingConfiguration2
 
class  TrainingConfiguration2List
 
class  TrainingProcess2
 
class  TrainingProcess2List
 
class  XmlElementFileDesc
 
class  XmlElementMLClass
 
class  XmlElementMLClassNameList
 Will only write the ClassName rather than complete MLClass instances. More...
 
class  XmlElementModel
 The base class to be used for the manufacturing if "Model" derived classes. More...
 
class  XmlElementModelParam
 Base class to be used by all ModelParam derived objects. More...
 
class  XmlElementModelParamTemplate
 ModelParam derived classes will implement their "XmlElement" helper class via this template. More...
 
class  XmlElementModelTemplate
 

Typedefs

typedef AttributeList const * AttributeListConstPtr
 
typedef AttributeList * AttributeListPtr
 
typedef Attribute * AttributePtr
 
typedef AttributeTypeVector * AttributeTypeVectorPtr
 
typedef BinaryClassParmsList::BinaryClassParmsListPtr BinaryClassParmsListPtr
 
typedef BinaryClassParms::BinaryClassParmsPtr BinaryClassParmsPtr
 
typedef ClassAssignments * ClassAssignmentsPtr
 
typedef ClassificationBiasMatrix::ClassificationBiasMatrixPtr ClassificationBiasMatrixPtr
 
typedef Classifier2List * Classifier2ListPtr
 
typedef Classifier2 * Classifier2Ptr
 
typedef ClassProbList * ClassProbListPtr
 
typedef ClassProb * ClassProbPtr
 
typedef ClassStatisticList * ClassStatisticListPtr
 
typedef ClassStatistic * ClassStatisticPtr
 
typedef ConfusionMatrix2::ConfusionMatrix2Ptr ConfusionMatrix2Ptr
 
typedef ConfussionMatrix2List * ConfussionMatrix2ListPtr
 
typedef CrossValidation * CrossValidationPtr
 
typedef CrossValidationVoting * CrossValidationVotingPtr
 
typedef DuplicateImageList * DuplicateImageListPtr
 
typedef DuplicateImage * DuplicateImagePtr
 
typedef DuplicateImages * DuplicateImagesPtr
 
typedef std::vector< double > Dvector
 
typedef FactoryFVProducer * FactoryFVProducerPtr
 
typedef FeatureEncoder2 * FeatureEncoder2Ptr
 
typedef FeatureEncoder * FeatureEncoderPtr
 
typedef FeatureFileIOC45::FeatureFileIOC45Ptr FeatureFileIOC45Ptr
 
typedef FeatureFileIO * FeatureFileIOPtr
 
typedef FeatureNumList const FeatureNumListConst
 
typedef FeatureNumListConst * FeatureNumListConstPtr
 
typedef FeatureNumList * FeatureNumListPtr
 
typedef FeatureVectorList * FeatureVectorListPtr
 
typedef FeatureVectorProducer * FeatureVectorProducerPtr
 
typedef FeatureVector * FeatureVectorPtr
 
typedef FeWhatToDo * FeWhatToDoPtr
 
typedef FileDesc::FileDescConstPtr FileDescConstPtr
 
typedef FileDescList * FileDescListPtr
 
typedef FileDesc * FileDescPtr
 
typedef std::vector< float > Fvector
 
typedef GrayScaleImagesFVList::GrayScaleImagesFVListPtr GrayScaleImagesFVListPtr
 
typedef GrayScaleImagesFVProducer * GrayScaleImagesFVProducerPtr
 
typedef GrayScaleImagesFV::GrayScaleImagesFVPtr GrayScaleImagesFVPtr
 
typedef ImageFeaturesDataIndexed * ImageFeaturesDataIndexedPtr
 
typedef ImageFeaturesNameIndexed * ImageFeaturesNameIndexedPtr
 
typedef std::vector< kkint32 > Ivector
 
typedef MLClassIndexList::MLClassIndexListPtr MLClassIndexListPtr
 
typedef MLClassList * MLClassListPtr
 
typedef MLClass * MLClassPtr
 
typedef ModelDual::ModelDualPtr ModelDualPtr
 
typedef ModelKnn::ModelKnnPtr ModelKnnPtr
 
typedef ModelOldSVM::ModelOldSVMPtr ModelOldSVMPtr
 
typedef ModelParamDual * ModelParamDualPtr
 
typedef ModelParamKnn::ModelParamKnnPtr ModelParamKnnPtr
 
typedef ModelParamOldSVM * ModelParamOldSVMPtr
 
typedef ModelParam::ModelParamPtr ModelParamPtr
 
typedef ModelParamSvmBase::ModelParamSvmBasePtr ModelParamSvmBasePtr
 
typedef ModelParamUsfCasCor * ModelParamUsfCasCorPtr
 
typedef Model::ModelPtr ModelPtr
 
typedef ModelSvmBase::ModelSvmBasePtr ModelSvmBasePtr
 
typedef ModelUsfCasCor::ModelUsfCasCorPtr ModelUsfCasCorPtr
 
typedef NormalizationParms * NormalizationParmsPtr
 
typedef Orderings * OrderingsPtr
 
typedef RunLog * RunLogPtr
 
typedef SizeDistribution * SizeDistributionPtr
 
typedef SVMModel * SVMModelPtr
 
typedef SVMparam * SVMparamPtr
 
typedef TrainingClassList * TrainingClassListPtr
 
typedef TrainingClass * TrainingClassPtr
 
typedef TrainingConfiguration2::TrainingConfiguration2Const TrainingConfiguration2Const
 
typedef TrainingConfiguration2::TrainingConfiguration2ConstPtr TrainingConfiguration2ConstPtr
 
typedef TrainingConfiguration2List * TrainingConfiguration2ListPtr
 
typedef TrainingConfiguration2 * TrainingConfiguration2Ptr
 
typedef TrainingProcess2List * TrainingProcess2ListPtr
 
typedef TrainingProcess2 * TrainingProcess2Ptr
 
typedef XmlElementTemplate< Attribute > XmlElementAttribute
 
typedef XmlElementTemplate< AttributeList > XmlElementAttributeList
 
typedef XmlElementAttributeList * XmlElementAttributeListPtr
 
typedef XmlElementAttribute * XmlElementAttributePtr
 
typedef XmlElementTemplate< AttributeTypeVector > XmlElementAttributeTypeVector
 
typedef XmlElementAttributeTypeVector * XmlElementAttributeTypeVectorPtr
 
typedef XmlElementTemplate< BinaryClassParmsList > XmlElementBinaryClassParmsList
 
typedef XmlElementBinaryClassParmsList * XmlElementBinaryClassParmsListPtr
 
typedef XmlElementTemplate< ClassProbList > XmlElementClassProbList
 
typedef XmlElementClassProbList * XmlElementClassProbListPtr
 
typedef XmlElementTemplate< FeatureNumList > XmlElementFeatureNumList
 
typedef XmlElementFeatureNumList * XmlElementFeatureNumListPtr
 
typedef XmlElementFileDesc * XmlElementFileDescPtr
 
typedef XmlElementTemplate< MLClassIndexList > XmlElementMLClassIndexList
 
typedef XmlElementMLClassIndexList * XmlElementMLClassIndexListPtr
 
typedef XmlElementMLClassNameList * XmlElementMLClassNameListPtr
 
typedef XmlElementMLClass * XmlElementMLClassPtr
 
typedef XmlElementModelTemplate< ModelDual > XmlElementModelDual
 
typedef XmlElementModelDual * XmlElementModelDualPtr
 
typedef XmlElementModelTemplate< ModelKnn > XmlElementModelKnn
 
typedef XmlElementModelKnn * XmlElementModelKnnPtr
 
typedef XmlElementModelTemplate< ModelOldSVM > XmlElementModelOldSVM
 
typedef XmlElementModelOldSVM * XmlElementModelOldSVMPtr
 
typedef XmlElementModelParamTemplate< ModelParamDual > XmlElementModelParamDual
 
typedef XmlElementModelParamDual * XmlElementModelParamDualPtr
 
typedef XmlElementModelParamTemplate< ModelParamKnn > XmlElementModelParamKnn
 
typedef XmlElementModelParamKnn * XmlElementModelParamKnnPtr
 
typedef XmlElementModelParamTemplate< ModelParamOldSVM > XmlElementModelParamOldSVM
 
typedef XmlElementModelParamOldSVM * XmlElementModelParamOldSVMPtr
 
typedef XmlElementModelParam * XmlElementModelParamPtr
 
typedef XmlElementModelParamTemplate< ModelParamSvmBase > XmlElementModelParamSvmBase
 
typedef XmlElementModelParamSvmBase * XmlElementModelParamSvmBasePtr
 
typedef XmlElementModelParamTemplate< ModelParamUsfCasCor > XmlElementModelParamUsfCasCor
 
typedef XmlElementModelParamUsfCasCor * XmlElementModelParamUsfCasCorPtr
 
typedef XmlElementModel * XmlElementModelPtr
 
typedef XmlElementModelTemplate< ModelSvmBase > XmlElementModelSvmBase
 
typedef XmlElementModelSvmBase * XmlElementModelSvmBasePtr
 
typedef XmlElementModelTemplate< ModelUsfCasCor > XmlElementModelUsfCasCor
 
typedef XmlElementModelUsfCasCor * XmlElementModelUsfCasCorPtr
 
typedef XmlElementTemplate< NormalizationParms > XmlElementNormalizationParms
 
typedef XmlElementNormalizationParms * XmlElementNormalizationParmsPtr
 
typedef XmlElementTemplate< SVMModel > XmlElementSVMModel
 
typedef XmlElementSVMModel * XmlElementSVMModelPtr
 
typedef XmlElementTemplate< SVMparam > XmlElementSVMparam
 
typedef XmlElementSVMparam * XmlElementSVMparamPtr
 
typedef XmlElementTemplate< TrainingClass > XmlElementTrainingClass
 
typedef XmlElementTemplate< TrainingClassList > XmlElementTrainingClassList
 
typedef XmlElementTrainingClassList * XmlElementTrainingClassListPtr
 
typedef XmlElementTrainingClass * XmlElementTrainingClassPtr
 
typedef XmlElementTemplate< TrainingConfiguration2 > XmlElementTrainingConfiguration2
 
typedef XmlElementTrainingConfiguration2 * XmlElementTrainingConfiguration2Ptr
 
typedef XmlElementTemplate< TrainingProcess2 > XmlElementTrainingProcess2
 
typedef XmlElementTrainingProcess2 * XmlElementTrainingProcess2Ptr
 
typedef struct svm_node * XSpacePtr
 

Enumerations

enum  AttributeType : int {
  AttributeType::NULLAttribute, AttributeType::Ignore, AttributeType::Numeric, AttributeType::Nominal,
  AttributeType::Ordinal, AttributeType::Symbolic
}
 
enum  FeWhatToDo { FeWhatToDo::FeAsIs, FeWhatToDo::FeBinary, FeWhatToDo::FeScale }
 
enum  LearnType {
  LearnType::NORMAL, LearnType::BAGGING, LearnType::BOOSTING, LearnType::SUBSPACE,
  LearnType::SAMPLESV
}
 
enum  SVM_EncodingMethod : int { SVM_EncodingMethod::Null, SVM_EncodingMethod::NoEncoding, SVM_EncodingMethod::Binary, SVM_EncodingMethod::Scaled }
 
enum  SVM_KernalType : int { SVM_KernalType::Null, SVM_KernalType::Linear, SVM_KernalType::Polynomial, SVM_KernalType::RBF }
 
enum  SVM_MachineType : int {
  SVM_MachineType::Null, SVM_MachineType::OneVsOne, SVM_MachineType::OneVsAll, SVM_MachineType::BinaryCombos,
  SVM_MachineType::BoostSVM
}
 
enum  SVM_SelectionMethod : int { SVM_SelectionMethod::Null, SVM_SelectionMethod::Voting, SVM_SelectionMethod::Probability }
 

Functions

AttributeType AttributeTypeFromStr (const KKStr &s)
 
const KKStr & AttributeTypeToStr (AttributeType type)
 
void EncodeProblem (const struct svm_paramater &param, struct svm_problem &prob_in, struct svm_problem &prob_out)
 
SVM_EncodingMethod EncodingMethodFromStr (const KKStr &encodingMethodStr)
 
KKStr EncodingMethodToStr (SVM_EncodingMethod encodingMethod)
 
SVM_KernalType KernalTypeFromStr (const KKStr &kernalTypeStr)
 
KKStr KernalTypeToStr (SVM_KernalType kernalType)
 
SVM_MachineType MachineTypeFromStr (const KKStr &machineTypeStr)
 
KKStr MachineTypeToStr (SVM_MachineType machineType)
 
AttributeType operator++ (AttributeType zed)
 
std::ostream & operator<< (std::ostream &os, const FeatureNumList &features)
 
std::ostream & operator<< (std::ostream &os, const MLClassList &classList)
 
KKStr & operator<< (KKStr &str, const MLClassList &classList)
 
KKStr & operator<< (KKStr &left, TrainingConfiguration2::ModelTypes modelingMethod)
 
std::ostream & operator<< (std::ostream &os, TrainingConfiguration2::ModelTypes modelingMethod)
 
ostream & operator<< (ostream &os, const FeatureNumList &features)
 
bool PairCompareOperator (ProbNamePair l, ProbNamePair r)
 
SVM_SelectionMethod SelectionMethodFromStr (const KKStr &selectionMethodStr)
 
KKStr SelectionMethodToStr (SVM_SelectionMethod selectionMethod)
 
void SvmDestroyModel (struct SvmModel233 **subModel)
 
struct SvmModel233 ** SvmLoadModel (istream &f, RunLog &log)
 
void SvmPredictClass (SVMparam &svmParam, struct SvmModel233 **subModel, const struct svm_node *unknownClassFeatureData, kkint32 *votes, double *probabilities, kkint32 knownClass, kkint32 &predClass1, kkint32 &predClass2, kkint32 &predClass1Votes, kkint32 &predClass2Votes, double &predClass1Prob, double &predClass2Prob, double &probOfKnownClass, Ivector &winners, double **crossClassProbTable, double &breakTie)
 
void SvmPredictRaw (SvmModel233 **submodel, const svm_node *unKnownData, double &label, double &dist)
 
kkint32 SvmPredictTwoClass (const struct svm_parameter &param, SvmModel233 **submodel, const svm_node *unKnownData, kkint32 desired, double &dist, double &probability, kkint32 excludeSupportVectorIDX)
 
void SvmSaveModel (ostream &o, struct SvmModel233 **model)
 
struct SvmModel233 ** SvmTrainModel (const struct svm_parameter &param, struct svm_problem &subprob)
 

Variables

const char * FeatureDecriptions []
 
MLClassList globalClassList
 

Detailed Description

Namespace for all K^2 Machine Learning code.

This library provides data structures that support Machine Learning functionality.

A library of routines used for Machine Learning, building Classifiers, etc.

"Machine Learning Code"

"Feature Data"

There are several Feature Data formats supported. Each one has its own class that is derived from 'FeatureFileIO'. The description of the data is managed by 'FileDesc'. For each type of dataset there will exist on one instance of a FileDesc class.

See also
Attribute
FeatureNumList
FeatureFileIO
FeatureVector
FeatureVectorList
MLClass
BitReduction
FeatureEncoder
NormalizationParms
Orderings

"Learning Algorithms"

There are several learning algorithms implemented. The Learning algorithms are all sub-classed from 'Model' and their related parameters are all sub-classed from 'ModelParam'.

See also
Model
ModelKnn
ModelOldSVM
ModelSvmBase
ModelParam
TrainingClass
TrainingProcess2

"Analysis Routines"

See also
ConfusionMatrix2
CrossValidation
CrossValidationMxN
CrossValidationVoting
SizeDistribution
Author
Kurt Kramer * Date: 2009-04-16 *

There are several classes defined in this library but the central ones that should be understood are MLClass, FeatureVector, Model, FileDesc, and Attribute. there is support for reading and writing several feature Data file formats C45 probably being the most well known.

Typedef Documentation

typedef AttributeList const* KKMLL::AttributeListConstPtr

Definition at line 214 of file Attribute.h.

typedef AttributeList* KKMLL::AttributeListPtr

Definition at line 212 of file Attribute.h.

typedef Attribute* KKMLL::AttributePtr

Definition at line 156 of file Attribute.h.

typedef AttributeTypeVector* KKMLL::AttributeTypeVectorPtr

Definition at line 72 of file Attribute.h.

typedef BinaryClassParmsList * KKMLL::BinaryClassParmsListPtr

Definition at line 220 of file BinaryClassParms.h.

typedef BinaryClassParms * KKMLL::BinaryClassParmsPtr

Definition at line 120 of file BinaryClassParms.h.

typedef ClassAssignments * KKMLL::ClassAssignmentsPtr

Definition at line 90 of file ClassAssignments.h.

typedef ClassificationBiasMatrix::ClassificationBiasMatrixPtr KKMLL::ClassificationBiasMatrixPtr

Definition at line 148 of file ClassificationBiasMatrix.h.

typedef Classifier2List * KKMLL::Classifier2ListPtr

Definition at line 72 of file Classifier2.h.

typedef Classifier2 * KKMLL::Classifier2Ptr

Definition at line 284 of file Classifier2.h.

typedef ClassProbList::ClassProbListPtr KKMLL::ClassProbListPtr

Definition at line 30 of file Classifier2.h.

typedef ClassProb * KKMLL::ClassProbPtr

Definition at line 28 of file Classifier2.h.

typedef ClassStatisticList* KKMLL::ClassStatisticListPtr

Definition at line 67 of file ClassStatistic.h.

typedef ClassStatistic* KKMLL::ClassStatisticPtr

Definition at line 38 of file ClassStatistic.h.

typedef ConfusionMatrix2 * KKMLL::ConfusionMatrix2Ptr

Definition at line 326 of file ConfusionMatrix2.h.

typedef ConfussionMatrix2List* KKMLL::ConfussionMatrix2ListPtr

Definition at line 343 of file ConfusionMatrix2.h.

typedef CrossValidation * KKMLL::CrossValidationPtr

Definition at line 213 of file CrossValidation.h.

typedef CrossValidationVoting* KKMLL::CrossValidationVotingPtr

Definition at line 120 of file CrossValidationVoting.h.

typedef DuplicateImageList * KKMLL::DuplicateImageListPtr

Definition at line 25 of file DuplicateImages.h.

typedef DuplicateImage * KKMLL::DuplicateImagePtr

Definition at line 28 of file DuplicateImages.h.

typedef DuplicateImages* KKMLL::DuplicateImagesPtr

Definition at line 119 of file DuplicateImages.h.

typedef std::vector<double> KKMLL::Dvector

Definition at line 20 of file SvmWrapper.h.

typedef FactoryFVProducer * KKMLL::FactoryFVProducerPtr

Definition at line 75 of file Model.h.

typedef FeatureEncoder2::FeatureEncoder2Ptr KKMLL::FeatureEncoder2Ptr

Definition at line 30 of file Model.h.

typedef FeatureEncoder * KKMLL::FeatureEncoderPtr

Definition at line 37 of file SVMModel.h.

typedef FeatureFileIOC45::FeatureFileIOC45Ptr KKMLL::FeatureFileIOC45Ptr

Definition at line 113 of file FeatureFileIOC45.h.

typedef FeatureFileIO * KKMLL::FeatureFileIOPtr

Definition at line 45 of file FileDesc.h.

typedef FeatureNumList const KKMLL::FeatureNumListConst

Definition at line 253 of file FeatureNumList.h.

typedef FeatureNumListConst * KKMLL::FeatureNumListConstPtr

Definition at line 254 of file FeatureNumList.h.

typedef FeatureNumList * KKMLL::FeatureNumListPtr

Definition at line 51 of file FeatureNumList.h.

typedef FeatureVectorList * KKMLL::FeatureVectorListPtr

Definition at line 46 of file Model.h.

typedef FeatureVectorProducer::FeatureVectorProducerPtr KKMLL::FeatureVectorProducerPtr

Definition at line 24 of file FactoryFVProducer.h.

typedef FeatureVector * KKMLL::FeatureVectorPtr

Definition at line 44 of file Model.h.

typedef FeWhatToDo* KKMLL::FeWhatToDoPtr

Definition at line 41 of file FeatureEncoder.h.

typedef FileDesc *const KKMLL::FileDescConstPtr

Definition at line 307 of file FileDesc.h.

typedef FileDescList * KKMLL::FileDescListPtr

Definition at line 41 of file FileDesc.h.

typedef FileDesc * KKMLL::FileDescPtr

Definition at line 46 of file FeatureNumList.h.

typedef std::vector<float> KKMLL::Fvector

Definition at line 19 of file SvmWrapper.h.

typedef GrayScaleImagesFVList::GrayScaleImagesFVListPtr KKMLL::GrayScaleImagesFVListPtr

Definition at line 459 of file GrayScaleImagesFV.h.

typedef GrayScaleImagesFVProducer* KKMLL::GrayScaleImagesFVProducerPtr

Definition at line 160 of file GrayScaleImagesFVProducer.h.

typedef GrayScaleImagesFV::GrayScaleImagesFVPtr KKMLL::GrayScaleImagesFVPtr

Definition at line 106 of file GrayScaleImagesFV.h.

typedef ImageFeaturesDataIndexed * KKMLL::ImageFeaturesDataIndexedPtr

Definition at line 33 of file DuplicateImages.h.

typedef ImageFeaturesNameIndexed * KKMLL::ImageFeaturesNameIndexedPtr

Definition at line 46 of file DuplicateImages.h.

typedef std::vector<kkint32> KKMLL::Ivector

Definition at line 18 of file SvmWrapper.h.

typedef MLClassIndexList * KKMLL::MLClassIndexListPtr

Definition at line 534 of file MLClass.h.

typedef MLClassList * KKMLL::MLClassListPtr

Definition at line 49 of file MLClass.h.

typedef MLClass * KKMLL::MLClassPtr

Definition at line 46 of file MLClass.h.

typedef ModelDual::ModelDualPtr KKMLL::ModelDualPtr

Definition at line 233 of file ModelDual.h.

typedef ModelKnn::ModelKnnPtr KKMLL::ModelKnnPtr

Definition at line 120 of file ModelKnn.h.

typedef ModelOldSVM * KKMLL::ModelOldSVMPtr

Definition at line 266 of file ModelOldSVM.h.

typedef ModelParamDual::ModelParamDualPtr KKMLL::ModelParamDualPtr

Definition at line 56 of file ModelDual.h.

typedef ModelParamKnn::ModelParamKnnPtr KKMLL::ModelParamKnnPtr

Definition at line 66 of file ModelParamKnn.h.

typedef ModelParamOldSVM * KKMLL::ModelParamOldSVMPtr

Definition at line 24 of file ModelOldSVM.h.

typedef ModelParam * KKMLL::ModelParamPtr

Definition at line 201 of file ModelParam.h.

typedef ModelParamSvmBase::ModelParamSvmBasePtr KKMLL::ModelParamSvmBasePtr

Definition at line 85 of file ModelParamSvmBase.h.

typedef ModelParamUsfCasCor* KKMLL::ModelParamUsfCasCorPtr

Definition at line 91 of file ModelParamUsfCasCor.h.

typedef Model::ModelPtr KKMLL::ModelPtr

Definition at line 443 of file Model.h.

typedef ModelSvmBase::ModelSvmBasePtr KKMLL::ModelSvmBasePtr

Definition at line 147 of file ModelSvmBase.h.

typedef ModelUsfCasCor::ModelUsfCasCorPtr KKMLL::ModelUsfCasCorPtr

Definition at line 141 of file ModelUsfCasCor.h.

typedef NormalizationParms * KKMLL::NormalizationParmsPtr

Definition at line 68 of file Model.h.

typedef Orderings::OrderingsPtr KKMLL::OrderingsPtr

Definition at line 58 of file CrossValidationMxN.h.

typedef RunLog * KKMLL::RunLogPtr

Definition at line 44 of file SVMModel.h.

typedef SizeDistribution* KKMLL::SizeDistributionPtr

Definition at line 77 of file SizeDistribution.h.

typedef SVMModel* KKMLL::SVMModelPtr

Definition at line 557 of file SVMModel.h.

typedef SVMparam* KKMLL::SVMparamPtr

Definition at line 286 of file SVMparam.h.

typedef TrainingClassList* KKMLL::TrainingClassListPtr

Definition at line 188 of file TrainingClass.h.

typedef TrainingClass* KKMLL::TrainingClassPtr

Definition at line 139 of file TrainingClass.h.

typedef TrainingConfiguration2::TrainingConfiguration2Const KKMLL::TrainingConfiguration2Const

Definition at line 552 of file TrainingConfiguration2.h.

typedef TrainingConfiguration2::TrainingConfiguration2ConstPtr KKMLL::TrainingConfiguration2ConstPtr

Definition at line 553 of file TrainingConfiguration2.h.

typedef TrainingConfiguration2List * KKMLL::TrainingConfiguration2ListPtr

Definition at line 50 of file TrainingConfiguration2.h.

typedef TrainingConfiguration2 * KKMLL::TrainingConfiguration2Ptr

Definition at line 56 of file Classifier2.h.

typedef TrainingProcess2List * KKMLL::TrainingProcess2ListPtr

summary>Manages the creation and loading of training models.

remarks> A trained model can either be built from scratch using specified training data or a persistent instance can be loaded from a XMLStream. There are several static methods that are best used to manage the various situations such as "CreateTrainingProcess", "CreateTrainingProcessForLevel", "CreateTrainingProcessFromTrainingExamples", "LoadExistingTrainingProcess". The result of any of these methods is a TrainigProcess2 instance that you use to create a Classifier instance.

Supporting Classes:

  • TrainingConfiguration2 Manages defines the parameters of a classifier such as:
    1. Type of algorithm
    2. List Classes of and where training examples can be loaded from.
    3. Sub-Classifiers that is for any given class another "TrainingConfiguration2" can be specified indicating another classifier to be used when that class is predicted.
    4. Class Weights

Model Base class for all algorithms. SupportVectorMachine(SVM), USFCasCor, BFS-SVM, etc ....

Definition at line 68 of file Classifier2.h.

typedef TrainingProcess2 * KKMLL::TrainingProcess2Ptr

Definition at line 62 of file Classifier2.h.

typedef XmlElementTemplate<Attribute> KKMLL::XmlElementAttribute

Definition at line 226 of file Attribute.h.

typedef XmlElementTemplate<AttributeList> KKMLL::XmlElementAttributeList

Definition at line 230 of file Attribute.h.

typedef XmlElementAttributeList* KKMLL::XmlElementAttributeListPtr

Definition at line 231 of file Attribute.h.

typedef XmlElementAttribute* KKMLL::XmlElementAttributePtr

Definition at line 227 of file Attribute.h.

typedef XmlElementTemplate<AttributeTypeVector> KKMLL::XmlElementAttributeTypeVector

Definition at line 222 of file Attribute.h.

typedef XmlElementAttributeTypeVector* KKMLL::XmlElementAttributeTypeVectorPtr

Definition at line 223 of file Attribute.h.

typedef XmlElementTemplate<BinaryClassParmsList> KKMLL::XmlElementBinaryClassParmsList

Definition at line 222 of file BinaryClassParms.h.

typedef XmlElementBinaryClassParmsList* KKMLL::XmlElementBinaryClassParmsListPtr

Definition at line 223 of file BinaryClassParms.h.

typedef XmlElementTemplate<ClassProbList> KKMLL::XmlElementClassProbList

Definition at line 146 of file ClassProb.h.

typedef XmlElementClassProbList* KKMLL::XmlElementClassProbListPtr

Definition at line 147 of file ClassProb.h.

typedef XmlElementTemplate<FeatureNumList> KKMLL::XmlElementFeatureNumList

Definition at line 263 of file FeatureNumList.h.

typedef XmlElementFeatureNumList* KKMLL::XmlElementFeatureNumListPtr

Definition at line 264 of file FeatureNumList.h.

typedef XmlElementFileDesc* KKMLL::XmlElementFileDescPtr

Definition at line 337 of file FileDesc.h.

typedef XmlElementTemplate<MLClassIndexList> KKMLL::XmlElementMLClassIndexList

Definition at line 606 of file MLClass.h.

typedef XmlElementMLClassIndexList* KKMLL::XmlElementMLClassIndexListPtr

Definition at line 607 of file MLClass.h.

typedef XmlElementMLClassNameList* KKMLL::XmlElementMLClassNameListPtr

Definition at line 603 of file MLClass.h.

typedef XmlElementMLClass* KKMLL::XmlElementMLClassPtr

Definition at line 572 of file MLClass.h.

typedef XmlElementModelTemplate<ModelDual> KKMLL::XmlElementModelDual

Definition at line 236 of file ModelDual.h.

typedef XmlElementModelDual* KKMLL::XmlElementModelDualPtr

Definition at line 237 of file ModelDual.h.

typedef XmlElementModelTemplate<ModelKnn> KKMLL::XmlElementModelKnn

Definition at line 123 of file ModelKnn.h.

typedef XmlElementModelKnn* KKMLL::XmlElementModelKnnPtr

Definition at line 124 of file ModelKnn.h.

typedef XmlElementModelTemplate<ModelOldSVM> KKMLL::XmlElementModelOldSVM

Definition at line 269 of file ModelOldSVM.h.

typedef XmlElementModelOldSVM* KKMLL::XmlElementModelOldSVMPtr

Definition at line 270 of file ModelOldSVM.h.

typedef XmlElementModelParamTemplate<ModelParamDual> KKMLL::XmlElementModelParamDual

Definition at line 104 of file ModelParamDual.h.

typedef XmlElementModelParamDual* KKMLL::XmlElementModelParamDualPtr

Definition at line 105 of file ModelParamDual.h.

typedef XmlElementModelParamTemplate<ModelParamKnn> KKMLL::XmlElementModelParamKnn

Definition at line 68 of file ModelParamKnn.h.

typedef XmlElementModelParamKnn* KKMLL::XmlElementModelParamKnnPtr

Definition at line 69 of file ModelParamKnn.h.

typedef XmlElementModelParamTemplate<ModelParamOldSVM> KKMLL::XmlElementModelParamOldSVM

Definition at line 185 of file ModelParamOldSVM.h.

typedef XmlElementModelParamOldSVM* KKMLL::XmlElementModelParamOldSVMPtr

Definition at line 186 of file ModelParamOldSVM.h.

typedef XmlElementModelParam* KKMLL::XmlElementModelParamPtr

Definition at line 248 of file ModelParam.h.

typedef XmlElementModelParamTemplate<ModelParamSvmBase> KKMLL::XmlElementModelParamSvmBase

Definition at line 87 of file ModelParamSvmBase.h.

typedef XmlElementModelParamSvmBase* KKMLL::XmlElementModelParamSvmBasePtr

Definition at line 88 of file ModelParamSvmBase.h.

typedef XmlElementModelParamTemplate<ModelParamUsfCasCor> KKMLL::XmlElementModelParamUsfCasCor

Definition at line 93 of file ModelParamUsfCasCor.h.

typedef XmlElementModelParamUsfCasCor* KKMLL::XmlElementModelParamUsfCasCorPtr

Definition at line 94 of file ModelParamUsfCasCor.h.

typedef XmlElementModel* KKMLL::XmlElementModelPtr

Definition at line 480 of file Model.h.

typedef XmlElementModelTemplate<ModelSvmBase> KKMLL::XmlElementModelSvmBase

Definition at line 150 of file ModelSvmBase.h.

typedef XmlElementModelSvmBase* KKMLL::XmlElementModelSvmBasePtr

Definition at line 151 of file ModelSvmBase.h.

typedef XmlElementModelTemplate<ModelUsfCasCor> KKMLL::XmlElementModelUsfCasCor

Definition at line 146 of file ModelUsfCasCor.h.

typedef XmlElementModelUsfCasCor* KKMLL::XmlElementModelUsfCasCorPtr

Definition at line 147 of file ModelUsfCasCor.h.

typedef XmlElementTemplate<NormalizationParms> KKMLL::XmlElementNormalizationParms

Definition at line 155 of file NormalizationParms.h.

typedef XmlElementNormalizationParms* KKMLL::XmlElementNormalizationParmsPtr

Definition at line 156 of file NormalizationParms.h.

typedef XmlElementTemplate<SVMModel> KKMLL::XmlElementSVMModel

Definition at line 561 of file SVMModel.h.

typedef XmlElementSVMModel* KKMLL::XmlElementSVMModelPtr

Definition at line 562 of file SVMModel.h.

typedef XmlElementTemplate<SVMparam> KKMLL::XmlElementSVMparam

Definition at line 302 of file SVMparam.h.

typedef XmlElementSVMparam* KKMLL::XmlElementSVMparamPtr

Definition at line 303 of file SVMparam.h.

typedef XmlElementTemplate<TrainingClass> KKMLL::XmlElementTrainingClass

Definition at line 192 of file TrainingClass.h.

typedef XmlElementTemplate<TrainingClassList> KKMLL::XmlElementTrainingClassList

Definition at line 196 of file TrainingClass.h.

typedef XmlElementTrainingClassList* KKMLL::XmlElementTrainingClassListPtr

Definition at line 197 of file TrainingClass.h.

typedef XmlElementTrainingClass* KKMLL::XmlElementTrainingClassPtr

Definition at line 193 of file TrainingClass.h.

typedef XmlElementTemplate<TrainingConfiguration2> KKMLL::XmlElementTrainingConfiguration2

Definition at line 587 of file TrainingConfiguration2.h.

typedef XmlElementTrainingConfiguration2* KKMLL::XmlElementTrainingConfiguration2Ptr

Definition at line 588 of file TrainingConfiguration2.h.

typedef XmlElementTemplate<TrainingProcess2> KKMLL::XmlElementTrainingProcess2

Definition at line 426 of file TrainingProcess2.h.

typedef XmlElementTrainingProcess2* KKMLL::XmlElementTrainingProcess2Ptr

Definition at line 427 of file TrainingProcess2.h.

typedef struct svm_node * KKMLL::XSpacePtr

Definition at line 58 of file SVMModel.h.

Enumeration Type Documentation

enum KKMLL::AttributeType : int
strong
Enumerator
NULLAttribute 
Ignore 
Numeric 
Nominal 
Ordinal 
Symbolic 

Same as Nominal, except the names file does not list all possible values. They have to be determined from the data file.

Definition at line 36 of file Attribute.h.

36  : int
37  {
38  NULLAttribute,
39  Ignore,
40  Numeric,
41  Nominal,
42  Ordinal,
43  Symbolic /**< Same as Nominal, except the names file does not
44  * list all possible values. They have to be determined from
45  * the data file.
46  */
47  };
enum KKMLL::FeWhatToDo
strong
Enumerator
FeAsIs 
FeBinary 
FeScale 

Definition at line 39 of file FeatureEncoder.h.

39 {FeAsIs, FeBinary, FeScale};
enum KKMLL::LearnType
strong
Enumerator
NORMAL 
BAGGING 
BOOSTING 
SUBSPACE 
SAMPLESV 

Definition at line 9 of file SvmWrapper.h.

10  {
11  NORMAL,
12  BAGGING,
13  BOOSTING,
14  SUBSPACE,
15  SAMPLESV
16  };
enum KKMLL::SVM_EncodingMethod : int
strong
Enumerator
Null 
NoEncoding 
Binary 
Scaled 

Definition at line 46 of file SVMparam.h.

46  : int
47  {Null,
48  NoEncoding,
49  Binary,
50  Scaled
51  };
enum KKMLL::SVM_KernalType : int
strong
Enumerator
Null 
Linear 
Polynomial 
RBF 

Definition at line 54 of file SVMparam.h.

54  : int
55  {Null,
56  Linear,
57  Polynomial,
58  RBF
59  };
enum KKMLL::SVM_MachineType : int
strong
Enumerator
Null 
OneVsOne 
OneVsAll 
BinaryCombos 
BoostSVM 

Definition at line 26 of file SVMparam.h.

26  : int
27  {Null,
28  OneVsOne,
29  OneVsAll,
30  BinaryCombos,
31  BoostSVM
32  };
enum KKMLL::SVM_SelectionMethod : int
strong
Enumerator
Null 
Voting 
Probability 

Definition at line 34 of file SVMparam.h.

34  : int
35  {Null,
36  Voting,
37  Probability
38  };

Function Documentation

AttributeType KKMLL::AttributeTypeFromStr ( const KKStr &  s)

Definition at line 442 of file Attribute.cpp.

References KKB::KKStr::EqualIgnoreCase(), Ignore, Nominal, NULLAttribute, Numeric, Ordinal, and Symbolic.

Referenced by KKMLL::Attribute::ReadXML().

443 {
444  if (s.EqualIgnoreCase ("Ignore"))
445  return AttributeType::Ignore;
446 
447  else if (s.EqualIgnoreCase ("Numeric"))
448  return AttributeType::Numeric;
449 
450  else if (s.EqualIgnoreCase ("Nominal"))
451  return AttributeType::Nominal;
452 
453  else if (s.EqualIgnoreCase ("Ordinal"))
454  return AttributeType::Ordinal;
455 
456  else if (s.EqualIgnoreCase ("Symbolic"))
457  return AttributeType::Symbolic;
458 
459  else
460  return AttributeType::NULLAttribute;
461 }
KKB::KKStr::EqualIgnoreCase
bool EqualIgnoreCase(const KKStr &s2) const
Definition: KKStr.cpp:1250
const KKStr & KKMLL::AttributeTypeToStr ( AttributeType  type)

Definition at line 422 of file Attribute.cpp.

Referenced by KKMLL::FeatureVectorList::FeatureTypeStr(), and KKMLL::Attribute::TypeStr().

423 {
424  static vector<KKStr> AttributeTypeStrings =
425  {
426  "Null",
427  "Ignore",
428  "Numeric",
429  "Nominal",
430  "Ordinal"
431  };
432 
433  if ((type < (AttributeType)0) || ((kkuint32)type >= AttributeTypeStrings.size ()))
434  return KKStr::EmptyStr ();
435 
436  return AttributeTypeStrings[(int)type];
437 } /* TypeStr */
KKB::kkuint32
unsigned __int32 kkuint32
Definition: KKBaseTypes.h:89
KKMLL::AttributeType
AttributeType
Definition: Attribute.h:36
void KKMLL::EncodeProblem ( const struct svm_paramater &  param,
struct svm_problem &  prob_in,
struct svm_problem &  prob_out 
)

Referenced by KKMLL::GrayScaleImagesFVProducer::DefineFileDescStatic(), and KKMLL::Attribute::WriteXML().

SVM_EncodingMethod KKMLL::EncodingMethodFromStr ( const KKStr &  encodingMethodStr)

Definition at line 899 of file SVMparam.cpp.

References Binary, KKB::KKStr::Concat(), NoEncoding, Null, KKB::KKStr::operator==(), Scaled, and KKB::KKStr::ToUpper().

Referenced by KKMLL::SVMparam::ParseCmdLineParameter(), and KKMLL::FeatureEncoder::ReadXML().

900 {
901  KKStr encodingMethodUpper = encodingMethodStr.ToUpper ();
902 
903  if ((encodingMethodUpper == "BINARY") || (encodingMethodUpper == "BIN"))
904  return SVM_EncodingMethod::Binary;
905 
906  if (encodingMethodUpper == "SCALE")
907  return SVM_EncodingMethod::Scaled;
908 
909  if (encodingMethodUpper == "NONE")
910  return SVM_EncodingMethod::NoEncoding;
911 
912  return SVM_EncodingMethod::Null;
913 } /* EncodingMethodFromStr */
KKB::KKStr::ToUpper
KKStr ToUpper() const
Definition: KKStr.cpp:2517
KKStr KKMLL::EncodingMethodToStr ( SVM_EncodingMethod  encodingMethod)

Referenced by KKMLL::SVMparam::ToString(), KKMLL::FeatureEncoder::WriteXML(), and KKMLL::SVMparam::WriteXML().

SVM_KernalType KKMLL::KernalTypeFromStr ( const KKStr &  kernalTypeStr)

Definition at line 932 of file SVMparam.cpp.

References KKB::KKStr::Concat(), Linear, KKB::KKStr::operator==(), Polynomial, RBF, and KKB::KKStr::Upper().

933 {
934  KKStr kernalTypeUpper = kernalTypeStr;
935  kernalTypeUpper.Upper ();
936 
937  if ((kernalTypeUpper == "LINEAR") ||
938  (kernalTypeUpper == "LIN") ||
939  (kernalTypeUpper == "L")
940  )
941  return SVM_KernalType::Linear;
942 
943  if ((kernalTypeUpper == "POLYNOMIAL") ||
944  (kernalTypeUpper == "POLY") ||
945  (kernalTypeUpper == "P") ||
946  (kernalTypeUpper == "PN")
947  )
948  return SVM_KernalType::Polynomial;
949 
950 
951  if ((kernalTypeUpper == "RBF") ||
952  (kernalTypeUpper == "R") ||
953  (kernalTypeUpper == "RADIALBASIS") ||
954  (kernalTypeUpper == "RADIALBASISFUNC") ||
955  (kernalTypeUpper == "RADIALBASISFUNCTION")
956  )
957  return SVM_KernalType::RBF;
958 
959  return SVM_KernalType::Linear;
960 } /* KernalTypeToStr */
KKB::KKStr::Upper
void Upper()
Converts all characters in string to their Upper case equivalents via &#39;toupper&#39;.
Definition: KKStr.cpp:2461
KKStr KKMLL::KernalTypeToStr ( SVM_KernalType  kernalType)

Definition at line 917 of file SVMparam.cpp.

References Linear, Polynomial, and RBF.

918 {
919  switch (kernalType)
920  {
921  case SVM_KernalType::Linear: return "Linear";
922  case SVM_KernalType::Polynomial: return "Polynomial";
923  case SVM_KernalType::RBF: return "RBF";
924  default: return "UnKnown";
925  }
926 
927  return "";
928 } /* KernalTypeToStr */
SVM_MachineType KKMLL::MachineTypeFromStr ( const KKStr &  machineTypeStr)

Definition at line 986 of file SVMparam.cpp.

References BinaryCombos, BoostSVM, KKB::KKStr::Concat(), Null, OneVsAll, OneVsOne, KKB::KKStr::operator==(), and KKB::KKStr::ToUpper().

Referenced by KKMLL::SVMparam::ParseCmdLineParameter().

987 {
988  KKStr machineTypeUpper = machineTypeStr.ToUpper ();
989 
990  if (machineTypeUpper == "ONEVSONE")
991  return SVM_MachineType::OneVsOne;
992 
993  else if (machineTypeUpper == "ONEVSALL")
994  return SVM_MachineType::OneVsAll;
995 
996  else if (machineTypeUpper == "BINARY")
997  return SVM_MachineType::BinaryCombos;
998 
999  else if (machineTypeUpper == "BOOSTSVM")
1000  return SVM_MachineType::BoostSVM;
1001 
1002  return SVM_MachineType::Null;
1003 } /* MachineTypeFromStr */
KKB::KKStr::ToUpper
KKStr ToUpper() const
Definition: KKStr.cpp:2517
KKStr KKMLL::MachineTypeToStr ( SVM_MachineType  machineType)

Definition at line 965 of file SVMparam.cpp.

References BinaryCombos, BoostSVM, OneVsAll, and OneVsOne.

Referenced by KKMLL::ModelOldSVM::Description(), KKMLL::SVMparam::ToString(), and KKMLL::SVMparam::WriteXML().

966 {
967  if (machineType == SVM_MachineType::OneVsOne)
968  return "OneVsOne";
969 
970  else if (machineType == SVM_MachineType::OneVsAll)
971  return "OneVsAll";
972 
973  else if (machineType == SVM_MachineType::BinaryCombos)
974  return "Binary";
975 
976  else if (machineType == SVM_MachineType::BoostSVM)
977  return "BoostSVM";
978 
979  return "";
980 } /* MachineTypeToStr */
AttributeType KKMLL::operator++ ( AttributeType  zed)

Definition at line 566 of file Attribute.cpp.

567 {
568  return (AttributeType)((int)zed + 1);
569 } /* operator++ */
KKMLL::AttributeType
AttributeType
Definition: Attribute.h:36
std::ostream & KKMLL::operator<< ( std::ostream &  os,
const FeatureNumList &  features 
)
std::ostream& KKMLL::operator<< ( std::ostream &  os,
const MLClassList &  classList 
)
KKStr & KKMLL::operator<< ( KKStr &  str,
const MLClassList &  classList 
)

Definition at line 1427 of file MLClass.cpp.

References KKMLL::MLClassList::ToString().

1430 {
1431  str << classList.ToString ();
1432  return str;
1433 }
KKMLL::MLClassList::ToString
KKStr ToString() const
Definition: MLClass.cpp:937
KKStr & KKMLL::operator<< ( KKStr &  left,
TrainingConfiguration2::ModelTypes  modelingMethod 
)

Definition at line 2002 of file TrainingConfiguration2.cpp.

References KKB::KKStr::Append(), and KKMLL::TrainingConfiguration2::ModelTypeToStr().

2005  {
2006  left.Append (TrainingConfiguration2::ModelTypeToStr (modelingMethod));
2007  return left;
2008  }
KKB::KKStr::Append
void Append(const char *buff)
Definition: KKStr.cpp:1783
ostream & KKMLL::operator<< ( std::ostream &  os,
TrainingConfiguration2::ModelTypes  modelingMethod 
)

Definition at line 2012 of file TrainingConfiguration2.cpp.

References KKMLL::TrainingConfiguration2::ModelTypeToStr(), and KKB::operator<<().

2015  {
2016  os << TrainingConfiguration2::ModelTypeToStr (modelingMethod);
2017  return os;
2018  }
ostream & KKMLL::operator<< ( ostream &  os,
const FeatureNumList &  features 
)

Definition at line 761 of file FeatureNumList.cpp.

References KKB::operator<<(), and KKMLL::FeatureNumList::ToString().

764  {
765  os << features.ToString ();
766  return os;
767  }
KKMLL::FeatureNumList::ToString
KKStr ToString() const
Returns comma delimited list of all features selected; will make use of range specification.
Definition: FeatureNumList.cpp:422
bool KKMLL::PairCompareOperator ( ProbNamePair  l,
ProbNamePair  r 
)

Definition at line 1568 of file SVMModel.cpp.

References KKMLL::ProbNamePair::probability.

1569  {
1570  return l.probability > r.probability;
1571  }
KKMLL::ProbNamePair::probability
double probability
Definition: SVMModel.h:72
SVM_SelectionMethod KKMLL::SelectionMethodFromStr ( const KKStr &  selectionMethodStr)

Definition at line 1022 of file SVMparam.cpp.

References KKB::KKStr::Concat(), Null, KKB::KKStr::operator==(), Probability, KKB::KKStr::ToUpper(), and Voting.

Referenced by KKMLL::SVMparam::ParseCmdLineParameter().

1023 {
1024  KKStr selectionMethodUpper = selectionMethodStr.ToUpper ();
1025 
1026  if ((selectionMethodUpper == "VOTE") || (selectionMethodUpper == "VOTING"))
1027  return SVM_SelectionMethod::Voting;
1028 
1029  if ((selectionMethodUpper == "P") ||
1030  (selectionMethodUpper == "PROB") ||
1031  (selectionMethodUpper == "PROBABILITY")
1032  )
1033  return SVM_SelectionMethod::Probability;
1034 
1035  return SVM_SelectionMethod::Null;
1036 
1037 } /* SelectionMethodFromStr */
KKB::KKStr::ToUpper
KKStr ToUpper() const
Definition: KKStr.cpp:2517
KKStr KKMLL::SelectionMethodToStr ( SVM_SelectionMethod  selectionMethod)

Definition at line 1008 of file SVMparam.cpp.

References Probability, and Voting.

Referenced by KKMLL::ModelOldSVM::Description(), KKMLL::SVMparam::ToString(), and KKMLL::SVMparam::WriteXML().

1009 {
1010  if (selectionMethod == SVM_SelectionMethod::Voting)
1011  return "Voting";
1012 
1013  if (selectionMethod == SVM_SelectionMethod::Probability)
1014  return "Probability";
1015 
1016  return "";
1017 } /* SelectionMethodToStr */
void KKMLL::SvmDestroyModel ( struct SvmModel233 **  subModel)

Definition at line 742 of file SvmWrapper.cpp.

References SVM233::svm_destroy_model().

743 {
744  svm_destroy_model (subModel[0]);
745 }
SVM233::svm_destroy_model
void svm_destroy_model(struct SvmModel233 *model)
Definition: svm.cpp:4442
struct SvmModel233 ** KKMLL::SvmLoadModel ( istream &  f,
RunLog &  log 
)

Definition at line 722 of file SvmWrapper.cpp.

References SVM233::Svm_Load_Model().

725 {
726  SvmModel233** models = new SvmModel233*[1];
727  models[0] = Svm_Load_Model (f, log);
728 
729  if (models[0] == NULL)
730  {
731  delete[] models;
732  models = NULL;
733  }
734 
735  return models;
736 }
SVM233::Svm_Load_Model
struct SvmModel233 * Svm_Load_Model(std::istream &f, RunLog &log)
void KKMLL::SvmPredictClass ( SVMparam &  svmParam,
struct SvmModel233 **  subModel,
const struct svm_node *  unknownClassFeatureData,
kkint32 *  votes,
double *  probabilities,
kkint32  knownClass,
kkint32 &  predClass1,
kkint32 &  predClass2,
kkint32 &  predClass1Votes,
kkint32 &  predClass2Votes,
double &  predClass1Prob,
double &  predClass2Prob,
double &  probOfKnownClass,
Ivector &  winners,
double **  crossClassProbTable,
double &  breakTie 
)
Parameters
[in]svmParamStructure that has all the parameters used for building the SVM.
[in]subModelThis is the model(classifier) returned by SvmTrainModel.
[in]probabilitiesArray that must be as big as number of classes the probability of each class will be returned where there sum is 1.0
[in]unknownClassFeatureDatadata structure you build that represents a sparse array of the feature data that is to be used for prediction.
[out]probabilitiesarray that will receive the predicted probability for each class.
[in]knownClassIf you happen to know hat the class really is you can specify it here so as to get the probability of it returned back to you.
[out]predClass1The prediction will be returned back in this field.
[out]predClass2The second most likely class based off either probability or number of votes.
[out]predClass1VotesNumber votes that class 'predClass1' received.
[out]predClass2VotesNumber votes that class 'predClass2' received.
[out]predClass1ProbProbability that class 'predClass1' received.
[out]predClass2ProbProbability that class 'predClass2' received.
[out]probOfKnownClassProbability that 'knownClass' class receved.
[out]winnersIf voting was specified in 'svmParam' and there was a tie between 2 or more classes; them the classes that ties will be in this vector.
[in]crossClassProbTableA 2 dimensional table that will have the computed probabilities between all the possible 2 class combinations.
[in]breakTieThe difference in probability between the two most likely classes.

Definition at line 585 of file SvmWrapper.cpp.

References GreaterVotes(), SVM233::SvmModel233::nr_class, KKMLL::SVMparam::Param(), Probability, and KKMLL::SVMparam::SelectionMethod().

602 {
603  const struct svm_parameter& param = svmParam.Param ();
604 
605  kkint32 NUMCLASS = subModel[0][0].nr_class;
606 
607  kkint32 numBinary = (NUMCLASS * (NUMCLASS - 1)) / 2;
608  Dvector dist (numBinary, 0);
609 
610 
611  svm_predict (subModel[0], unknownClassFeatureData, dist, winners, -1);
612 
613 
614  ComputeProb (NUMCLASS,
615  svmParam.ProbClassPairs (),
616  dist, // Distances for each binary classifier from decision boundary.
617  crossClassProbTable, // Will get Probabilities between classes.
618  votes,
619  probabilities, // Probabilities for Each Class
620  knownClass // -1 = Don't know the class otherwise the Number of the Class.
621  );
622 
623  GreaterVotes ((svmParam.SelectionMethod () == SVM_SelectionMethod::Probability),
624  NUMCLASS,
625  votes,
626  probabilities,
627  predClass1,
628  predClass2
629  );
630  if (predClass1 >= 0)
631  {
632  predClass1Votes = votes[predClass1];
633  predClass1Prob = probabilities[predClass1];
634  }
635 
636  if (predClass2 >= 0)
637  {
638  predClass2Votes = votes[predClass2];
639  predClass2Prob = probabilities[predClass2];
640  }
641 
642  if (knownClass >= 0)
643  probOfKnownClass = probabilities[knownClass];
644 
645  breakTie = (predClass1Prob - predClass2Prob);
646 } /* SvmPredictClass */
KKMLL::Dvector
std::vector< double > Dvector
Definition: SvmWrapper.h:20
KKB::kkint32
__int32 kkint32
Definition: KKBaseTypes.h:88
ComputeProb
void ComputeProb(kkint32 numClasses, const VectorFloat &probClassPairs, vector< double > &dist, double **crossClassProbTable, kkint32 *votes, double *probabilities, kkint32 knownClassNum)
Definition: SvmWrapper.cpp:472
KKMLL::SVMparam::ProbClassPairs
const VectorFloat & ProbClassPairs() const
Definition: SVMparam.h:161
SVM233::SvmModel233::nr_class
kkint32 nr_class
Definition: svm.h:138
GreaterVotes
void GreaterVotes(bool useProbability, kkint32 numClasses, kkint32 *votes, double *probabilities, kkint32 &pred1Idx, kkint32 &pred2Idx)
Definition: SvmWrapper.cpp:175
SVM233::svm_predict
double svm_predict(const struct SvmModel233 *model, const struct svm_node *x)
KKMLL::SVMparam::SelectionMethod
SVM_SelectionMethod SelectionMethod() const
Definition: SVMparam.h:169
SVM233::svm_parameter
Definition: svm.h:65
KKMLL::SVMparam::Param
const svm_parameter & Param() const
Definition: SVMparam.h:159
void KKMLL::SvmPredictRaw ( SvmModel233 **  submodel,
const svm_node *  unKnownData,
double &  label,
double &  dist 
)

Definition at line 684 of file SvmWrapper.cpp.

References SVM233::SvmModel233::nr_class, KKB::osWaitForEnter(), and SVM233::svm_predictTwoClasses().

Referenced by KKMLL::SVMModel::PredictRaw().

689 {
690  if (submodel[0]->nr_class != 2)
691  {
692  cerr << endl
693  << endl
694  << "SvmPredictTwoClass *** ERROR ***" << endl
695  << endl
696  << "Number of classes should be equal to two." << endl
697  << endl;
698  osWaitForEnter ();
699  exit (-1);
700  }
701 
702  dist = 0.0;
703 
704  label = svm_predictTwoClasses (submodel[0], unKnownData, dist, -1);
705 
706  return;
707 } /* SvmPredictRaw */
KKB::endl
HTMLReport &__cdecl endl(HTMLReport &htmlReport)
Definition: HTMLReport.cpp:240
KKB::osWaitForEnter
void osWaitForEnter()
Definition: OSservices.cpp:1513
SVM233::svm_predictTwoClasses
double svm_predictTwoClasses(const SvmModel233 *model, const svm_node *x, double &dist, kkint32 excludeSupportVectorIDX)
Definition: svm.cpp:4218
kkint32 KKMLL::SvmPredictTwoClass ( const struct svm_parameter &  param,
SvmModel233 **  submodel,
const svm_node *  unKnownData,
kkint32  desired,
double &  dist,
double &  probability,
kkint32  excludeSupportVectorIDX 
)
void KKMLL::SvmSaveModel ( ostream &  o,
struct SvmModel233 **  model 
)

Definition at line 711 of file SvmWrapper.cpp.

References SVM233::Svm_Save_Model().

714 {
715  Svm_Save_Model (o, model[0]);
716 }
SVM233::Svm_Save_Model
void Svm_Save_Model(std::ostream &o, const SvmModel233 *model)
struct SvmModel233 ** KKMLL::SvmTrainModel ( const struct svm_parameter &  param,
struct svm_problem &  subprob 
)

Definition at line 543 of file SvmWrapper.cpp.

References BAGGING, BOOSTING, SVM233::svm_parameter::boosting, SVM233::svm_parameter::dimSelect, NORMAL, SVM233::svm_parameter::numSVM, SVM233::svm_parameter::sample, SAMPLESV, SVM233::svm_parameter::sampleSV, SUBSPACE, and SVM233::svm_train().

546 {
547  struct SvmModel233 **submodel;
548 
549  kkint32 numSVM = param.numSVM;
550  kkint32 sample = (kkint32) (param.sample);
551 
552  //kkint32 numClass=param.numClass;
553  kkint32 sampleSV = param.sampleSV;
554  kkint32 boosting = param.boosting;
555  kkint32 dimSelect = param.dimSelect;
556 
557  LearnType learnType;
558 
559  if ((numSVM == 1) && (sample == 100))
560  learnType= LearnType::NORMAL;
561 
562  else if (dimSelect > 0)
563  learnType = LearnType::SUBSPACE;
564 
565  else if (boosting != 0)
566  learnType=LearnType::BOOSTING;
567 
568  else if(sampleSV!=0)
569  learnType=LearnType::SAMPLESV;
570 
571  else
572  learnType=LearnType::BAGGING;
573 
574  submodel = new SvmModel233* [numSVM];
575  submodel[0] = svm_train (&subprob, &param);
576 
577  return submodel;
578 } /* SvmTrainModel */
SVM233::svm_parameter::dimSelect
kkint32 dimSelect
Definition: svm.h:102
KKB::kkint32
__int32 kkint32
Definition: KKBaseTypes.h:88
SVM233::svm_parameter::boosting
kkint32 boosting
Definition: svm.h:98
KKMLL::LearnType
LearnType
Definition: SvmWrapper.h:9
SVM233::svm_parameter::sampleSV
kkint32 sampleSV
Definition: svm.h:108
SVM233::svm_train
struct SvmModel233 * svm_train(const struct svm_problem *prob, const struct svm_parameter *param)
SVM233::svm_parameter::sample
float sample
Definition: svm.h:107
SVM233::svm_parameter::numSVM
kkint32 numSVM
Definition: svm.h:106

Variable Documentation

const char * KKMLL::FeatureDecriptions

Definition at line 314 of file FeatureNumList_Old.h.

MLClassList KKMLL::globalClassList